wehub-resource-sync
5144 lines
440 KiB
Plaintext
5144 lines
440 KiB
Plaintext
{
|
|
"cells": [
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "19ddb599",
|
|
"metadata": {
|
|
"papermill": {
|
|
"duration": 0.001881,
|
|
"end_time": "2026-03-24T02:22:15.050003+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:22:15.048122+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"source": [
|
|
"# ETFs: Feature Engineering\n",
|
|
"\n",
|
|
"This notebook implements feature engineering for the ETFs case study.\n",
|
|
"Features are organized around the core hypothesis: cross-asset momentum\n",
|
|
"combined with regime conditioning.\n",
|
|
"\n",
|
|
"## Learning Objectives\n",
|
|
"\n",
|
|
"- Compute risk-adjusted momentum scores including skip-recent (12-1)\n",
|
|
"- Construct cross-sectional ranks for primary horizon features only\n",
|
|
"- Implement regime indicators (yield curve slope, SPY-TLT correlation)\n",
|
|
"- Enforce point-in-time eligibility on the feature matrix\n",
|
|
"- Diagnose feature quality via autocorrelation and rank stability\n",
|
|
"\n",
|
|
"## Book Reference\n",
|
|
"\n",
|
|
"Chapter 8, Section 8.2 (Price-Derived Features)\n",
|
|
"\n",
|
|
"## Prerequisites\n",
|
|
"\n",
|
|
"- [`01_feasibility_analysis`](01_feasibility_analysis.ipynb) (produces `eligibility.csv`)\n",
|
|
"- [`02_labels`](02_labels.ipynb) (produces label parquet files and `config/cv_config.json`)"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 1,
|
|
"id": "f87a91de",
|
|
"metadata": {
|
|
"execution": {
|
|
"iopub.execute_input": "2026-04-09T01:05:21.955441Z",
|
|
"iopub.status.busy": "2026-04-09T01:05:21.955324Z",
|
|
"iopub.status.idle": "2026-04-09T01:05:27.427490Z",
|
|
"shell.execute_reply": "2026-04-09T01:05:27.426792Z"
|
|
},
|
|
"papermill": {
|
|
"duration": 1.298855,
|
|
"end_time": "2026-03-24T02:22:16.350759+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:22:15.051904+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"outputs": [
|
|
{
|
|
"name": "stderr",
|
|
"output_type": "stream",
|
|
"text": [
|
|
".venv/lib/python3.14/site-packages/ml4t/engineer/features/ml/__init__.py:9: UserWarning: Feature 'cyclical_encode': lookback=0 but has period/window parameter. Consider using lookback='period' or specifying the actual lookback.\n",
|
|
" from ml4t.engineer.features.ml.cyclical_encode import * # noqa: F403\n"
|
|
]
|
|
}
|
|
],
|
|
"source": [
|
|
"\"\"\"ETFs: Feature Engineering.\"\"\"\n",
|
|
"\n",
|
|
"import itertools\n",
|
|
"import os\n",
|
|
"import warnings\n",
|
|
"\n",
|
|
"import numpy as np\n",
|
|
"import plotly.graph_objects as go\n",
|
|
"import polars as pl\n",
|
|
"from ml4t.diagnostic.evaluation.stats import benjamini_hochberg_fdr\n",
|
|
"from ml4t.diagnostic.metrics import compute_ic_hac_stats\n",
|
|
"from ml4t.engineer.features.momentum import adx, aroon, cci, macd, rsi, stochastic\n",
|
|
"from ml4t.engineer.features.regime import choppiness_index, hurst_exponent\n",
|
|
"from ml4t.engineer.features.trend import ema, sma\n",
|
|
"from ml4t.engineer.features.volatility import natr\n",
|
|
"from ml4t.engineer.features.volume import obv\n",
|
|
"from scipy.stats import spearmanr\n",
|
|
"\n",
|
|
"from data import load_etfs, load_macro\n",
|
|
"from utils.paths import get_case_study_dir\n",
|
|
"\n",
|
|
"warnings.filterwarnings(\"ignore\")\n",
|
|
"\n",
|
|
"CASE_DIR = get_case_study_dir(\"etfs\")\n",
|
|
"FEATURES_DIR = CASE_DIR / \"features\"\n",
|
|
"\n",
|
|
"NUMERIC_DTYPES = {\n",
|
|
" pl.Boolean,\n",
|
|
" pl.Int8,\n",
|
|
" pl.Int16,\n",
|
|
" pl.Int32,\n",
|
|
" pl.Int64,\n",
|
|
" pl.UInt8,\n",
|
|
" pl.UInt16,\n",
|
|
" pl.UInt32,\n",
|
|
" pl.UInt64,\n",
|
|
" pl.Float32,\n",
|
|
" pl.Float64,\n",
|
|
"}\n",
|
|
"\n",
|
|
"\n",
|
|
"def as_float(value: object | None) -> float | None:\n",
|
|
" \"\"\"Convert Polars scalar outputs to plain float for plotting.\"\"\"\n",
|
|
" if value is None:\n",
|
|
" return None\n",
|
|
" return float(str(value))"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "8bc23671",
|
|
"metadata": {
|
|
"tags": [
|
|
"parameters"
|
|
]
|
|
},
|
|
"outputs": [],
|
|
"source": [
|
|
"# Production defaults — Papermill injects overrides for CI"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "b37bcab6",
|
|
"metadata": {
|
|
"papermill": {
|
|
"duration": 0.001972,
|
|
"end_time": "2026-03-24T02:22:16.354869+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:22:16.352897+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"source": [
|
|
"## Configuration"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 2,
|
|
"id": "ad5f486f",
|
|
"metadata": {
|
|
"execution": {
|
|
"iopub.execute_input": "2026-04-09T01:05:27.429520Z",
|
|
"iopub.status.busy": "2026-04-09T01:05:27.429371Z",
|
|
"iopub.status.idle": "2026-04-09T01:05:27.432107Z",
|
|
"shell.execute_reply": "2026-04-09T01:05:27.431401Z"
|
|
},
|
|
"papermill": {
|
|
"duration": 0.004421,
|
|
"end_time": "2026-03-24T02:22:16.361222+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:22:16.356801+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"outputs": [],
|
|
"source": [
|
|
"# Feature parameters\n",
|
|
"REGIME_THRESHOLD = 0.005 # 0.5% yield curve slope\n",
|
|
"\n",
|
|
"# Multi-horizon lookback periods (trading days)\n",
|
|
"MOMENTUM_HORIZONS = [5, 10, 21, 42, 63, 126, 189, 252]\n",
|
|
"VOLATILITY_HORIZONS = [21, 63, 126, 252]"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "9af2ae73",
|
|
"metadata": {
|
|
"papermill": {
|
|
"duration": 0.001888,
|
|
"end_time": "2026-03-24T02:22:16.365050+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:22:16.363162+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"source": [
|
|
"## 1. Load Data"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 3,
|
|
"id": "02f6d515",
|
|
"metadata": {
|
|
"execution": {
|
|
"iopub.execute_input": "2026-04-09T01:05:27.433347Z",
|
|
"iopub.status.busy": "2026-04-09T01:05:27.433262Z",
|
|
"iopub.status.idle": "2026-04-09T01:05:27.513795Z",
|
|
"shell.execute_reply": "2026-04-09T01:05:27.513060Z"
|
|
},
|
|
"papermill": {
|
|
"duration": 0.056674,
|
|
"end_time": "2026-03-24T02:22:16.423617+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:22:16.366943+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"outputs": [
|
|
{
|
|
"name": "stdout",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"ETF Universe: 100 assets, 470,662 rows\n"
|
|
]
|
|
}
|
|
],
|
|
"source": [
|
|
"prices = (\n",
|
|
" load_etfs()\n",
|
|
" .select([\"symbol\", \"timestamp\", \"open\", \"high\", \"low\", \"close\", \"volume\"])\n",
|
|
" .sort([\"symbol\", \"timestamp\"])\n",
|
|
")\n",
|
|
"\n",
|
|
"# Load eligibility for PIT filtering\n",
|
|
"eligibility = pl.read_csv(CASE_DIR / \"eligibility.csv\")\n",
|
|
"if \"symbol\" in eligibility.columns and \"symbol\" not in eligibility.columns:\n",
|
|
" eligibility = eligibility\n",
|
|
"\n",
|
|
"assets_list = sorted(prices[\"symbol\"].unique().to_list())\n",
|
|
"n_assets = prices[\"symbol\"].n_unique()\n",
|
|
"print(f\"ETF Universe: {n_assets} assets, {len(prices):,} rows\")"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "147971b6",
|
|
"metadata": {
|
|
"papermill": {
|
|
"duration": 0.00322,
|
|
"end_time": "2026-03-24T02:22:16.430720+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:22:16.427500+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"source": [
|
|
"### Yield Curve Data"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 4,
|
|
"id": "cfc407ac",
|
|
"metadata": {
|
|
"execution": {
|
|
"iopub.execute_input": "2026-04-09T01:05:27.515459Z",
|
|
"iopub.status.busy": "2026-04-09T01:05:27.515299Z",
|
|
"iopub.status.idle": "2026-04-09T01:05:27.540412Z",
|
|
"shell.execute_reply": "2026-04-09T01:05:27.539621Z"
|
|
},
|
|
"papermill": {
|
|
"duration": 0.008419,
|
|
"end_time": "2026-03-24T02:22:16.441472+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:22:16.433053+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"outputs": [
|
|
{
|
|
"name": "stdout",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"Yield curve data: 9,495 rows\n"
|
|
]
|
|
}
|
|
],
|
|
"source": [
|
|
"macro_data = load_macro()\n",
|
|
"\n",
|
|
"# Compute yield curve slope from dgs10 and dgs2\n",
|
|
"yield_curve = (\n",
|
|
" macro_data.select([\"timestamp\", \"dgs10\", \"dgs2\"])\n",
|
|
" .with_columns(\n",
|
|
" (pl.col(\"dgs10\") / 100).alias(\"yield_10y\"),\n",
|
|
" (pl.col(\"dgs2\") / 100).alias(\"yield_2y\"),\n",
|
|
" ((pl.col(\"dgs10\") - pl.col(\"dgs2\")) / 100).alias(\"slope\"),\n",
|
|
" )\n",
|
|
" .select([\"timestamp\", \"yield_10y\", \"yield_2y\", \"slope\"])\n",
|
|
" .drop_nulls()\n",
|
|
" .sort(\"timestamp\")\n",
|
|
")\n",
|
|
"\n",
|
|
"print(f\"Yield curve data: {len(yield_curve):,} rows\")"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "649dc630",
|
|
"metadata": {
|
|
"papermill": {
|
|
"duration": 0.001949,
|
|
"end_time": "2026-03-24T02:22:16.445564+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:22:16.443615+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"source": [
|
|
"## 2. Feature Engineering\n",
|
|
"\n",
|
|
"Features are organized into focused families aligned with the momentum +\n",
|
|
"regime hypothesis. We avoid feature inflation by ranking only primary\n",
|
|
"horizons and keeping a curated set of MA ratios."
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "392bf438",
|
|
"metadata": {
|
|
"lines_to_next_cell": 2,
|
|
"papermill": {
|
|
"duration": 0.001932,
|
|
"end_time": "2026-03-24T02:22:16.449458+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:22:16.447526+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"source": [
|
|
"### 2.1 Momentum and Returns\n",
|
|
"\n",
|
|
"Multi-horizon raw returns, skip-recent momentum (12-1 and 6-1), risk-adjusted\n",
|
|
"momentum (Sharpe-like), and momentum acceleration."
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 5,
|
|
"id": "1c6ddc96",
|
|
"metadata": {
|
|
"execution": {
|
|
"iopub.execute_input": "2026-04-09T01:05:27.541949Z",
|
|
"iopub.status.busy": "2026-04-09T01:05:27.541798Z",
|
|
"iopub.status.idle": "2026-04-09T01:05:27.547268Z",
|
|
"shell.execute_reply": "2026-04-09T01:05:27.546629Z"
|
|
},
|
|
"lines_to_next_cell": 2,
|
|
"papermill": {
|
|
"duration": 0.005976,
|
|
"end_time": "2026-03-24T02:22:16.457392+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:22:16.451416+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"outputs": [],
|
|
"source": [
|
|
"def compute_momentum_features(df: pl.DataFrame) -> pl.DataFrame:\n",
|
|
" \"\"\"Momentum, volatility, and risk-adjusted returns.\"\"\"\n",
|
|
" df = df.sort([\"symbol\", \"timestamp\"])\n",
|
|
"\n",
|
|
" # Multi-horizon raw returns (clip(1e-8) guards against div-by-zero → inf)\n",
|
|
" momentum_cols = [\n",
|
|
" (\n",
|
|
" pl.col(\"close\") / pl.col(\"close\").shift(h).over(\"symbol\").clip(lower_bound=1e-8) - 1\n",
|
|
" ).alias(f\"ret_{h}d\")\n",
|
|
" for h in MOMENTUM_HORIZONS\n",
|
|
" ]\n",
|
|
" df = df.with_columns(momentum_cols)\n",
|
|
"\n",
|
|
" # Skip-recent momentum (Jegadeesh and Titman 1993)\n",
|
|
" # 12-1: 252d return excluding most recent 21d\n",
|
|
" # 6-1: 126d return excluding most recent 21d\n",
|
|
" df = df.with_columns(\n",
|
|
" (pl.col(\"ret_252d\") - pl.col(\"ret_21d\")).alias(\"skip_recent_12_1\"),\n",
|
|
" (pl.col(\"ret_126d\") - pl.col(\"ret_21d\")).alias(\"skip_recent_6_1\"),\n",
|
|
" )\n",
|
|
"\n",
|
|
" # Log returns for volatility\n",
|
|
" df = df.with_columns(pl.col(\"close\").log().diff().over(\"symbol\").alias(\"log_return\"))\n",
|
|
"\n",
|
|
" # Multi-horizon volatility (annualized)\n",
|
|
" volatility_cols = [\n",
|
|
" pl.col(\"log_return\").rolling_std(h).over(\"symbol\").mul(np.sqrt(252)).alias(f\"vol_{h}d\")\n",
|
|
" for h in VOLATILITY_HORIZONS\n",
|
|
" ]\n",
|
|
" df = df.with_columns(volatility_cols)\n",
|
|
"\n",
|
|
" # Volatility-adjusted momentum (Sharpe-like)\n",
|
|
" sharpe_cols = [\n",
|
|
" (\n",
|
|
" pl.col(\"log_return\").rolling_sum(h).over(\"symbol\")\n",
|
|
" / pl.col(\"log_return\").rolling_std(h).over(\"symbol\")\n",
|
|
" * np.sqrt(252 / h)\n",
|
|
" ).alias(f\"sharpe_{h}d\")\n",
|
|
" for h in MOMENTUM_HORIZONS\n",
|
|
" ]\n",
|
|
" df = df.with_columns(sharpe_cols)\n",
|
|
"\n",
|
|
" # Momentum acceleration\n",
|
|
" df = df.with_columns(\n",
|
|
" (pl.col(\"ret_21d\") - pl.col(\"ret_63d\")).alias(\"mom_accel_short\"),\n",
|
|
" (pl.col(\"ret_63d\") - pl.col(\"ret_126d\")).alias(\"mom_accel_medium\"),\n",
|
|
" (pl.col(\"ret_126d\") - pl.col(\"ret_252d\")).alias(\"mom_accel_long\"),\n",
|
|
" )\n",
|
|
"\n",
|
|
" # Volatility ratios (short/long)\n",
|
|
" df = df.with_columns(\n",
|
|
" (pl.col(\"vol_21d\") / pl.col(\"vol_63d\")).alias(\"vol_ratio_short\"),\n",
|
|
" (pl.col(\"vol_63d\") / pl.col(\"vol_126d\")).alias(\"vol_ratio_medium\"),\n",
|
|
" )\n",
|
|
"\n",
|
|
" return df"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "b6ad44b8",
|
|
"metadata": {
|
|
"lines_to_next_cell": 2,
|
|
"papermill": {
|
|
"duration": 0.001973,
|
|
"end_time": "2026-03-24T02:22:16.461358+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:22:16.459385+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"source": [
|
|
"### 2.2 Technical Indicators\n",
|
|
"\n",
|
|
"Curated set: oscillators (RSI, MACD, ADX, CCI, Stochastic, Aroon),\n",
|
|
"select MA ratios (SMA 50/200, EMA 26 only -- review recommended dropping\n",
|
|
"the rest), Bollinger %B, and NATR."
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 6,
|
|
"id": "d3e82b85",
|
|
"metadata": {
|
|
"execution": {
|
|
"iopub.execute_input": "2026-04-09T01:05:27.548517Z",
|
|
"iopub.status.busy": "2026-04-09T01:05:27.548419Z",
|
|
"iopub.status.idle": "2026-04-09T01:05:27.552272Z",
|
|
"shell.execute_reply": "2026-04-09T01:05:27.551785Z"
|
|
},
|
|
"lines_to_next_cell": 2,
|
|
"papermill": {
|
|
"duration": 0.005628,
|
|
"end_time": "2026-03-24T02:22:16.468920+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:22:16.463292+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"outputs": [],
|
|
"source": [
|
|
"def compute_technical_features(df: pl.DataFrame) -> pl.DataFrame:\n",
|
|
" \"\"\"Technical indicators and trend features.\"\"\"\n",
|
|
" # Oscillators\n",
|
|
" df = df.with_columns(\n",
|
|
" rsi(\"close\", period=7).over(\"symbol\").alias(\"rsi_7\"),\n",
|
|
" rsi(\"close\", period=14).over(\"symbol\").alias(\"rsi_14\"),\n",
|
|
" macd(\"close\", fast_period=12, slow_period=26).over(\"symbol\").alias(\"macd_line\"),\n",
|
|
" adx(\"high\", \"low\", \"close\", period=14).over(\"symbol\").alias(\"adx_14\"),\n",
|
|
" cci(\"high\", \"low\", \"close\", period=14).over(\"symbol\").alias(\"cci_14\"),\n",
|
|
" cci(\"high\", \"low\", \"close\", period=21).over(\"symbol\").alias(\"cci_21\"),\n",
|
|
" stochastic(\"high\", \"low\", \"close\", fastk_period=14).over(\"symbol\").alias(\"stoch_k\"),\n",
|
|
" )\n",
|
|
"\n",
|
|
" # Aroon oscillator\n",
|
|
" df = (\n",
|
|
" df.with_columns(aroon(\"high\", \"low\", timeperiod=25).over(\"symbol\").alias(\"_aroon_struct\"))\n",
|
|
" .with_columns(\n",
|
|
" (\n",
|
|
" pl.col(\"_aroon_struct\").struct.field(\"up\")\n",
|
|
" - pl.col(\"_aroon_struct\").struct.field(\"down\")\n",
|
|
" ).alias(\"aroon_diff\")\n",
|
|
" )\n",
|
|
" .drop(\"_aroon_struct\")\n",
|
|
" )\n",
|
|
"\n",
|
|
" # MA ratios -- curated set (SMA 50/200, EMA 26 only)\n",
|
|
" df = df.with_columns(\n",
|
|
" (pl.col(\"close\") / sma(\"close\", period=50).over(\"symbol\")).alias(\"sma_ratio_50\"),\n",
|
|
" (pl.col(\"close\") / sma(\"close\", period=200).over(\"symbol\")).alias(\"sma_ratio_200\"),\n",
|
|
" (pl.col(\"close\") / ema(\"close\", period=26).over(\"symbol\")).alias(\"ema_ratio_26\"),\n",
|
|
" )\n",
|
|
"\n",
|
|
" # Bollinger %B\n",
|
|
" df = df.with_columns(\n",
|
|
" pl.col(\"close\").rolling_mean(20).over(\"symbol\").alias(\"_bb_mid\"),\n",
|
|
" pl.col(\"close\").rolling_std(20).over(\"symbol\").alias(\"_bb_std\"),\n",
|
|
" )\n",
|
|
" df = df.with_columns(\n",
|
|
" (\n",
|
|
" (pl.col(\"close\") - (pl.col(\"_bb_mid\") - 2 * pl.col(\"_bb_std\")))\n",
|
|
" / (4 * pl.col(\"_bb_std\"))\n",
|
|
" ).alias(\"bb_pctb_20\")\n",
|
|
" ).drop([\"_bb_mid\", \"_bb_std\"])\n",
|
|
"\n",
|
|
" # NATR\n",
|
|
" df = df.with_columns(\n",
|
|
" natr(\"high\", \"low\", \"close\", period=14).over(\"symbol\").alias(\"natr_14\"),\n",
|
|
" )\n",
|
|
"\n",
|
|
" return df"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "a52a50b8",
|
|
"metadata": {
|
|
"lines_to_next_cell": 2,
|
|
"papermill": {
|
|
"duration": 0.001923,
|
|
"end_time": "2026-03-24T02:22:16.472837+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:22:16.470914+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"source": [
|
|
"### 2.3 Regime, Risk, Volume, and Distance Features"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 7,
|
|
"id": "9aa6c3e2",
|
|
"metadata": {
|
|
"execution": {
|
|
"iopub.execute_input": "2026-04-09T01:05:27.554544Z",
|
|
"iopub.status.busy": "2026-04-09T01:05:27.553847Z",
|
|
"iopub.status.idle": "2026-04-09T01:05:27.571376Z",
|
|
"shell.execute_reply": "2026-04-09T01:05:27.570733Z"
|
|
},
|
|
"lines_to_next_cell": 2,
|
|
"papermill": {
|
|
"duration": 0.013025,
|
|
"end_time": "2026-03-24T02:22:16.487809+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:22:16.474784+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"outputs": [],
|
|
"source": [
|
|
"def compute_regime_volume_features(df: pl.DataFrame) -> pl.DataFrame:\n",
|
|
" \"\"\"Regime indicators, risk metrics, volume features, distance extremes.\"\"\"\n",
|
|
" # Regime indicators\n",
|
|
" df = df.with_columns(\n",
|
|
" choppiness_index(\"high\", \"low\", \"close\", period=14).over(\"symbol\").alias(\"chop_14\"),\n",
|
|
" hurst_exponent(\"close\", period=100).over(\"symbol\").alias(\"hurst_100\"),\n",
|
|
" )\n",
|
|
"\n",
|
|
" # Maximum drawdown\n",
|
|
" df = df.with_columns(\n",
|
|
" (\n",
|
|
" (pl.col(\"close\") - pl.col(\"close\").rolling_max(63).over(\"symbol\"))\n",
|
|
" / pl.col(\"close\").rolling_max(63).over(\"symbol\").clip(lower_bound=1e-8)\n",
|
|
" ).alias(\"max_dd_63d\"),\n",
|
|
" (\n",
|
|
" (pl.col(\"close\") - pl.col(\"close\").rolling_max(126).over(\"symbol\"))\n",
|
|
" / pl.col(\"close\").rolling_max(126).over(\"symbol\").clip(lower_bound=1e-8)\n",
|
|
" ).alias(\"max_dd_126d\"),\n",
|
|
" )\n",
|
|
"\n",
|
|
" # Volume features (winsorize relative volume at 1st/99th)\n",
|
|
" df = df.with_columns(\n",
|
|
" (\n",
|
|
" pl.col(\"volume\")\n",
|
|
" / pl.col(\"volume\").rolling_mean(21).over(\"symbol\").clip(lower_bound=1e-8)\n",
|
|
" ).alias(\"vol_ratio_21d_raw\"),\n",
|
|
" (\n",
|
|
" pl.col(\"volume\")\n",
|
|
" / pl.col(\"volume\").rolling_mean(63).over(\"symbol\").clip(lower_bound=1e-8)\n",
|
|
" ).alias(\"vol_ratio_63d_raw\"),\n",
|
|
" )\n",
|
|
" df = df.with_columns(\n",
|
|
" pl.col(\"vol_ratio_21d_raw\")\n",
|
|
" .clip(\n",
|
|
" pl.col(\"vol_ratio_21d_raw\").quantile(0.01).over(\"timestamp\"),\n",
|
|
" pl.col(\"vol_ratio_21d_raw\").quantile(0.99).over(\"timestamp\"),\n",
|
|
" )\n",
|
|
" .alias(\"vol_ratio_21d\"),\n",
|
|
" pl.col(\"vol_ratio_63d_raw\")\n",
|
|
" .clip(\n",
|
|
" pl.col(\"vol_ratio_63d_raw\").quantile(0.01).over(\"timestamp\"),\n",
|
|
" pl.col(\"vol_ratio_63d_raw\").quantile(0.99).over(\"timestamp\"),\n",
|
|
" )\n",
|
|
" .alias(\"vol_ratio_63d\"),\n",
|
|
" ).drop([\"vol_ratio_21d_raw\", \"vol_ratio_63d_raw\"])\n",
|
|
"\n",
|
|
" # OBV z-score\n",
|
|
" df = df.with_columns(obv(\"close\", \"volume\").over(\"symbol\").alias(\"_obv\"))\n",
|
|
" df = df.with_columns(\n",
|
|
" (\n",
|
|
" (pl.col(\"_obv\") - pl.col(\"_obv\").rolling_mean(63).over(\"symbol\"))\n",
|
|
" / pl.col(\"_obv\").rolling_std(63).over(\"symbol\")\n",
|
|
" ).alias(\"obv_zscore_63d\"),\n",
|
|
" ).drop(\"_obv\")\n",
|
|
"\n",
|
|
" # Consistency and distance from extremes\n",
|
|
" df = df.with_columns(\n",
|
|
" (pl.col(\"log_return\") > 0)\n",
|
|
" .cast(pl.Float64)\n",
|
|
" .rolling_mean(63)\n",
|
|
" .over(\"symbol\")\n",
|
|
" .alias(\"pct_positive_63d\"),\n",
|
|
" )\n",
|
|
" df = df.with_columns(\n",
|
|
" (\n",
|
|
" pl.col(\"close\") / pl.col(\"close\").rolling_max(252).over(\"symbol\").clip(lower_bound=1e-8)\n",
|
|
" ).alias(\"dist_52w_high\"),\n",
|
|
" (\n",
|
|
" pl.col(\"close\") / pl.col(\"close\").rolling_min(252).over(\"symbol\").clip(lower_bound=1e-8)\n",
|
|
" ).alias(\"dist_52w_low\"),\n",
|
|
" )\n",
|
|
"\n",
|
|
" return df"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "a606bcae",
|
|
"metadata": {
|
|
"lines_to_next_cell": 2,
|
|
"papermill": {
|
|
"duration": 0.002031,
|
|
"end_time": "2026-03-24T02:22:16.492033+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:22:16.490002+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"source": [
|
|
"### 2.4 Cross-Asset Correlation Feature\n",
|
|
"\n",
|
|
"Rolling 63-day correlation between SPY and TLT captures the equity-bond\n",
|
|
"relationship. When this flips from negative to positive, diversification\n",
|
|
"breaks down -- a valuable regime indicator for cross-asset rotation."
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 8,
|
|
"id": "a2c502a1",
|
|
"metadata": {
|
|
"execution": {
|
|
"iopub.execute_input": "2026-04-09T01:05:27.572988Z",
|
|
"iopub.status.busy": "2026-04-09T01:05:27.572885Z",
|
|
"iopub.status.idle": "2026-04-09T01:05:27.575817Z",
|
|
"shell.execute_reply": "2026-04-09T01:05:27.575252Z"
|
|
},
|
|
"papermill": {
|
|
"duration": 0.00474,
|
|
"end_time": "2026-03-24T02:22:16.498763+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:22:16.494023+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"outputs": [],
|
|
"source": [
|
|
"def compute_cross_asset_features(df: pl.DataFrame) -> pl.DataFrame:\n",
|
|
" \"\"\"Compute cross-asset correlation features (SPY vs TLT).\"\"\"\n",
|
|
" # Extract SPY and TLT log returns\n",
|
|
" spy_ret = (\n",
|
|
" df.filter(pl.col(\"symbol\") == \"SPY\")\n",
|
|
" .select([\"timestamp\", \"log_return\"])\n",
|
|
" .rename({\"log_return\": \"_spy_ret\"})\n",
|
|
" )\n",
|
|
" tlt_ret = (\n",
|
|
" df.filter(pl.col(\"symbol\") == \"TLT\")\n",
|
|
" .select([\"timestamp\", \"log_return\"])\n",
|
|
" .rename({\"log_return\": \"_tlt_ret\"})\n",
|
|
" )\n",
|
|
"\n",
|
|
" # Join SPY and TLT returns on date\n",
|
|
" corr_df = spy_ret.join(tlt_ret, on=\"timestamp\", how=\"inner\").sort(\"timestamp\")\n",
|
|
"\n",
|
|
" # Rolling 63-day correlation\n",
|
|
" corr_df = corr_df.with_columns(\n",
|
|
" pl.rolling_corr(pl.col(\"_spy_ret\"), pl.col(\"_tlt_ret\"), window_size=63).alias(\n",
|
|
" \"corr_spy_tlt_63d\"\n",
|
|
" )\n",
|
|
" ).select([\"timestamp\", \"corr_spy_tlt_63d\"])\n",
|
|
"\n",
|
|
" # Broadcast to all assets\n",
|
|
" return df.join(corr_df, on=\"timestamp\", how=\"left\")"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "78bd41a1",
|
|
"metadata": {
|
|
"papermill": {
|
|
"duration": 0.002089,
|
|
"end_time": "2026-03-24T02:22:16.502871+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:22:16.500782+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"source": [
|
|
"### Apply All Feature Groups"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 9,
|
|
"id": "5b590e00",
|
|
"metadata": {
|
|
"execution": {
|
|
"iopub.execute_input": "2026-04-09T01:05:27.576919Z",
|
|
"iopub.status.busy": "2026-04-09T01:05:27.576816Z",
|
|
"iopub.status.idle": "2026-04-09T01:06:26.774197Z",
|
|
"shell.execute_reply": "2026-04-09T01:06:26.773731Z"
|
|
},
|
|
"papermill": {
|
|
"duration": 27.939101,
|
|
"end_time": "2026-03-24T02:22:44.443963+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:22:16.504862+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"outputs": [
|
|
{
|
|
"name": "stdout",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"2026-04-08 21:05:28 | mlquant.features.adx | INFO | [ADX] Starting calculation with parameters: period=14 (shape: (4,))\n"
|
|
]
|
|
},
|
|
{
|
|
"name": "stdout",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"2026-04-08 21:05:28 | mlquant.features.adx | INFO | [ADX] Completed calculation (shape: (4,)) (0.05ms)\n"
|
|
]
|
|
},
|
|
{
|
|
"name": "stdout",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"Shape after feature computation: (470662, 59)\n"
|
|
]
|
|
}
|
|
],
|
|
"source": [
|
|
"features_raw = (\n",
|
|
" prices.pipe(compute_momentum_features)\n",
|
|
" .pipe(compute_technical_features)\n",
|
|
" .pipe(compute_regime_volume_features)\n",
|
|
" .pipe(compute_cross_asset_features)\n",
|
|
")\n",
|
|
"\n",
|
|
"print(f\"Shape after feature computation: {features_raw.shape}\")"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "776ce437",
|
|
"metadata": {
|
|
"papermill": {
|
|
"duration": 0.002104,
|
|
"end_time": "2026-03-24T02:22:44.448407+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:22:44.446303+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"source": [
|
|
"### 2.5 Cross-Sectional Ranking\n",
|
|
"\n",
|
|
"Rank only the primary horizon features to avoid multicollinearity from\n",
|
|
"ranking every horizon. Three rank features: `ret_126d_rank`,\n",
|
|
"`sharpe_126d_rank`, `vol_63d_rank`."
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 10,
|
|
"id": "3a336db8",
|
|
"metadata": {
|
|
"execution": {
|
|
"iopub.execute_input": "2026-04-09T01:06:26.775302Z",
|
|
"iopub.status.busy": "2026-04-09T01:06:26.775211Z",
|
|
"iopub.status.idle": "2026-04-09T01:06:26.840405Z",
|
|
"shell.execute_reply": "2026-04-09T01:06:26.839938Z"
|
|
},
|
|
"papermill": {
|
|
"duration": 0.042449,
|
|
"end_time": "2026-03-24T02:22:44.492914+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:22:44.450465+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"outputs": [],
|
|
"source": [
|
|
"# Rank only primary features (review recommendation)\n",
|
|
"RANK_FEATURES = [\"ret_126d\", \"sharpe_126d\", \"vol_63d\"]\n",
|
|
"\n",
|
|
"rank_exprs = [\n",
|
|
" (pl.col(col).rank().over(\"timestamp\") / pl.col(col).count().over(\"timestamp\")).alias(\n",
|
|
" f\"{col}_rank\"\n",
|
|
" )\n",
|
|
" for col in RANK_FEATURES\n",
|
|
"]\n",
|
|
"features_raw = features_raw.with_columns(rank_exprs)"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "55d2ebe8",
|
|
"metadata": {
|
|
"lines_to_next_cell": 2,
|
|
"papermill": {
|
|
"duration": 0.003153,
|
|
"end_time": "2026-03-24T02:22:44.498703+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:22:44.495550+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"source": [
|
|
"### 2.6 Regime Indicators\n",
|
|
"\n",
|
|
"Two regime features from the yield curve:\n",
|
|
"1. Binary regime (slope > 0.5%)\n",
|
|
"2. Continuous yield curve slope\n",
|
|
"3. Rolling z-score of slope (review recommendation)"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 11,
|
|
"id": "3a8fe789",
|
|
"metadata": {
|
|
"execution": {
|
|
"iopub.execute_input": "2026-04-09T01:06:26.841663Z",
|
|
"iopub.status.busy": "2026-04-09T01:06:26.841578Z",
|
|
"iopub.status.idle": "2026-04-09T01:06:26.916659Z",
|
|
"shell.execute_reply": "2026-04-09T01:06:26.916169Z"
|
|
},
|
|
"papermill": {
|
|
"duration": 0.052318,
|
|
"end_time": "2026-03-24T02:22:44.553789+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:22:44.501471+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"outputs": [
|
|
{
|
|
"name": "stdout",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"Regime Distribution: shape: (2, 2)\n",
|
|
"┌────────┬────────┐\n",
|
|
"│ regime ┆ count │\n",
|
|
"│ --- ┆ --- │\n",
|
|
"│ i32 ┆ u32 │\n",
|
|
"╞════════╪════════╡\n",
|
|
"│ 0 ┆ 168549 │\n",
|
|
"│ 1 ┆ 302113 │\n",
|
|
"└────────┴────────┘\n"
|
|
]
|
|
}
|
|
],
|
|
"source": [
|
|
"def add_regime_indicators(\n",
|
|
" df: pl.DataFrame, yield_curve: pl.DataFrame, threshold: float = 0.005\n",
|
|
") -> pl.DataFrame:\n",
|
|
" \"\"\"Add regime indicators from yield curve data.\"\"\"\n",
|
|
" regime_df = (\n",
|
|
" yield_curve.with_columns(\n",
|
|
" pl.when(pl.col(\"slope\") > threshold).then(1).otherwise(0).alias(\"regime\"),\n",
|
|
" pl.col(\"slope\").alias(\"yield_curve_slope\"),\n",
|
|
" # Rolling z-score of slope\n",
|
|
" (\n",
|
|
" (pl.col(\"slope\") - pl.col(\"slope\").rolling_mean(252))\n",
|
|
" / pl.col(\"slope\").rolling_std(252)\n",
|
|
" ).alias(\"yield_curve_zscore\"),\n",
|
|
" )\n",
|
|
" .select([\"timestamp\", \"regime\", \"yield_curve_slope\", \"yield_curve_zscore\"])\n",
|
|
" .sort(\"timestamp\")\n",
|
|
" )\n",
|
|
"\n",
|
|
" return df.sort(\"timestamp\").join_asof(regime_df, on=\"timestamp\", strategy=\"backward\")\n",
|
|
"\n",
|
|
"\n",
|
|
"features_raw = add_regime_indicators(features_raw, yield_curve, threshold=REGIME_THRESHOLD)\n",
|
|
"\n",
|
|
"regime_dist = features_raw.group_by(\"regime\").agg(pl.len().alias(\"count\"))\n",
|
|
"print(\"Regime Distribution:\", regime_dist)"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "b84e142f",
|
|
"metadata": {
|
|
"papermill": {
|
|
"duration": 0.002102,
|
|
"end_time": "2026-03-24T02:22:44.558246+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:22:44.556144+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"source": [
|
|
"### 2.7 Point-in-Time Eligibility Filter\n",
|
|
"\n",
|
|
"Join features to eligibility.csv to ensure features are only kept for\n",
|
|
"ETFs that were tradable at each date. This prevents inflation of the\n",
|
|
"cross-section breadth in early years when fewer ETFs met the ADV threshold."
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 12,
|
|
"id": "0a8ddc48",
|
|
"metadata": {
|
|
"execution": {
|
|
"iopub.execute_input": "2026-04-09T01:06:26.917749Z",
|
|
"iopub.status.busy": "2026-04-09T01:06:26.917597Z",
|
|
"iopub.status.idle": "2026-04-09T01:06:26.975502Z",
|
|
"shell.execute_reply": "2026-04-09T01:06:26.974974Z"
|
|
},
|
|
"papermill": {
|
|
"duration": 0.03723,
|
|
"end_time": "2026-03-24T02:22:44.597527+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:22:44.560297+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"outputs": [
|
|
{
|
|
"name": "stdout",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"Eligibility filter: 470,662 -> 396,186 rows (74,476 dropped)"
|
|
]
|
|
},
|
|
{
|
|
"name": "stdout",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"\n"
|
|
]
|
|
}
|
|
],
|
|
"source": [
|
|
"# Add year for eligibility join\n",
|
|
"features_raw = features_raw.with_columns(pl.col(\"timestamp\").dt.year().alias(\"_year\"))\n",
|
|
"\n",
|
|
"elig_pairs = eligibility.select(\n",
|
|
" pl.col(\"symbol\"),\n",
|
|
" pl.col(\"eligible_year\").alias(\"_year\"),\n",
|
|
")\n",
|
|
"\n",
|
|
"# Semi-join: keep only rows where (asset, year) exists in eligibility\n",
|
|
"features_eligible = features_raw.join(elig_pairs, on=[\"symbol\", \"_year\"], how=\"semi\").drop(\"_year\")\n",
|
|
"\n",
|
|
"n_dropped = len(features_raw) - len(features_eligible)\n",
|
|
"print(\n",
|
|
" f\"Eligibility filter: {len(features_raw):,} -> {len(features_eligible):,} rows ({n_dropped:,} dropped)\"\n",
|
|
")"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "e79a86b5",
|
|
"metadata": {
|
|
"papermill": {
|
|
"duration": 0.002133,
|
|
"end_time": "2026-03-24T02:22:44.601962+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:22:44.599829+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"source": [
|
|
"## 3. Prepare Final Feature Matrix"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 13,
|
|
"id": "82f958ac",
|
|
"metadata": {
|
|
"execution": {
|
|
"iopub.execute_input": "2026-04-09T01:06:26.977109Z",
|
|
"iopub.status.busy": "2026-04-09T01:06:26.976960Z",
|
|
"iopub.status.idle": "2026-04-09T01:06:27.075482Z",
|
|
"shell.execute_reply": "2026-04-09T01:06:27.074944Z"
|
|
},
|
|
"papermill": {
|
|
"duration": 0.053892,
|
|
"end_time": "2026-03-24T02:22:44.657918+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:22:44.604026+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"outputs": [
|
|
{
|
|
"name": "stdout",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"No features with >5% missing after warmup drop\n",
|
|
"\n",
|
|
"Feature Summary\n",
|
|
" Total features: 57\n",
|
|
" Rows: 396,186\n",
|
|
" Assets: 99\n",
|
|
" Date range: 2007-01-03 to 2025-12-31\n",
|
|
"\n",
|
|
"Feature categories:\n",
|
|
" Momentum (raw): 8\n",
|
|
" Skip-recent: 2\n",
|
|
" Sharpe: 8\n",
|
|
" Volatility: 5\n",
|
|
" Vol ratios: 4\n",
|
|
" Technical: 8\n",
|
|
" Trend (MA): 3\n",
|
|
" Regime/macro: 6\n",
|
|
" Risk (DD): 2\n",
|
|
" Ranks: 3\n",
|
|
" Other: 8\n"
|
|
]
|
|
}
|
|
],
|
|
"source": [
|
|
"# Columns to exclude from feature set\n",
|
|
"exclude_cols = {\n",
|
|
" \"timestamp\",\n",
|
|
" \"symbol\",\n",
|
|
" \"open\",\n",
|
|
" \"high\",\n",
|
|
" \"low\",\n",
|
|
" \"close\",\n",
|
|
" \"volume\",\n",
|
|
" \"log_return\",\n",
|
|
"}\n",
|
|
"\n",
|
|
"feature_cols = [c for c in features_eligible.columns if c not in exclude_cols]\n",
|
|
"non_numeric_feature_cols = [\n",
|
|
" c for c in feature_cols if features_eligible.schema[c] not in NUMERIC_DTYPES\n",
|
|
"]\n",
|
|
"feature_cols = [c for c in feature_cols if c not in non_numeric_feature_cols]\n",
|
|
"\n",
|
|
"assert \"log_return\" not in feature_cols, \"log_return must be excluded (contemporaneous with label)\"\n",
|
|
"unexpected_feature_keys = {\"timestamp\", \"symbol\"} & set(feature_cols)\n",
|
|
"assert not unexpected_feature_keys, (\n",
|
|
" f\"Feature matrix leaked key columns: {sorted(unexpected_feature_keys)}\"\n",
|
|
")\n",
|
|
"assert not non_numeric_feature_cols, (\n",
|
|
" f\"Non-numeric columns leaked into features: {non_numeric_feature_cols}\"\n",
|
|
")\n",
|
|
"required_core_features = {\n",
|
|
" \"ret_21d\",\n",
|
|
" \"sharpe_126d\",\n",
|
|
" \"rsi_14\",\n",
|
|
" \"ema_ratio_26\",\n",
|
|
" \"yield_curve_slope\",\n",
|
|
" \"corr_spy_tlt_63d\",\n",
|
|
" \"max_dd_63d\",\n",
|
|
"}\n",
|
|
"missing_core_features = required_core_features - set(feature_cols)\n",
|
|
"assert not missing_core_features, f\"Missing core curated features: {sorted(missing_core_features)}\"\n",
|
|
"\n",
|
|
"# Prepare final DataFrame\n",
|
|
"features = (\n",
|
|
" features_eligible.select([\"timestamp\", \"symbol\"] + feature_cols)\n",
|
|
" .drop_nulls(subset=[\"sharpe_126d\"]) # Drop warmup rows\n",
|
|
" .sort([\"timestamp\", \"symbol\"])\n",
|
|
")\n",
|
|
"\n",
|
|
"# Missing values check\n",
|
|
"high_missing = [\n",
|
|
" (c, features[c].null_count() / len(features))\n",
|
|
" for c in feature_cols\n",
|
|
" if features[c].null_count() / len(features) > 0.05\n",
|
|
"]\n",
|
|
"if high_missing:\n",
|
|
" print(\"Features with >5% missing after warmup drop:\")\n",
|
|
" for col, pct in high_missing:\n",
|
|
" print(f\" {col}: {pct:.1%}\")\n",
|
|
"else:\n",
|
|
" print(\"No features with >5% missing after warmup drop\")\n",
|
|
"\n",
|
|
"print(\"\\nFeature Summary\")\n",
|
|
"print(f\" Total features: {len(feature_cols)}\")\n",
|
|
"print(f\" Rows: {len(features):,}\")\n",
|
|
"print(f\" Assets: {features['symbol'].n_unique()}\")\n",
|
|
"print(f\" Date range: {features['timestamp'].min()} to {features['timestamp'].max()}\")\n",
|
|
"\n",
|
|
"# Feature category counts\n",
|
|
"cats = {\n",
|
|
" \"Momentum (raw)\": sum(\n",
|
|
" 1 for c in feature_cols if c.startswith(\"ret_\") and not c.endswith(\"_rank\")\n",
|
|
" ),\n",
|
|
" \"Skip-recent\": sum(1 for c in feature_cols if c.startswith(\"skip_\")),\n",
|
|
" \"Sharpe\": sum(1 for c in feature_cols if c.startswith(\"sharpe_\") and not c.endswith(\"_rank\")),\n",
|
|
" \"Volatility\": sum(\n",
|
|
" 1 for c in feature_cols if c.startswith(\"vol_\") and not c.startswith(\"vol_ratio\")\n",
|
|
" ),\n",
|
|
" \"Vol ratios\": sum(1 for c in feature_cols if c.startswith(\"vol_ratio\")),\n",
|
|
" \"Technical\": sum(\n",
|
|
" 1\n",
|
|
" for c in feature_cols\n",
|
|
" if any(c.startswith(p) for p in [\"rsi_\", \"macd\", \"adx_\", \"cci_\", \"stoch_\", \"aroon\"])\n",
|
|
" ),\n",
|
|
" \"Trend (MA)\": sum(1 for c in feature_cols if any(c.startswith(p) for p in [\"sma_\", \"ema_\"])),\n",
|
|
" \"Regime/macro\": sum(\n",
|
|
" 1\n",
|
|
" for c in feature_cols\n",
|
|
" if any(c.startswith(p) for p in [\"regime\", \"yield_\", \"chop_\", \"hurst_\", \"corr_spy\"])\n",
|
|
" ),\n",
|
|
" \"Risk (DD)\": sum(1 for c in feature_cols if c.startswith(\"max_dd\")),\n",
|
|
" \"Ranks\": sum(1 for c in feature_cols if c.endswith(\"_rank\")),\n",
|
|
" \"Other\": 0, # computed below\n",
|
|
"}\n",
|
|
"accounted = sum(cats.values())\n",
|
|
"cats[\"Other\"] = len(feature_cols) - accounted\n",
|
|
"\n",
|
|
"print(\"\\nFeature categories:\")\n",
|
|
"for cat, count in cats.items():\n",
|
|
" if count > 0:\n",
|
|
" print(f\" {cat}: {count}\")"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "4c771c6b",
|
|
"metadata": {
|
|
"papermill": {
|
|
"duration": 0.002123,
|
|
"end_time": "2026-03-24T02:22:44.662357+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:22:44.660234+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"source": [
|
|
"## 4. Feature Diagnostics"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "a4e3d39b",
|
|
"metadata": {
|
|
"papermill": {
|
|
"duration": 0.002019,
|
|
"end_time": "2026-03-24T02:22:44.666449+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:22:44.664430+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"source": [
|
|
"### 4.1 Autocorrelation (Signal Persistence)"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 14,
|
|
"id": "96c71dce",
|
|
"metadata": {
|
|
"execution": {
|
|
"iopub.execute_input": "2026-04-09T01:06:27.077169Z",
|
|
"iopub.status.busy": "2026-04-09T01:06:27.077052Z",
|
|
"iopub.status.idle": "2026-04-09T01:06:27.639137Z",
|
|
"shell.execute_reply": "2026-04-09T01:06:27.638381Z"
|
|
},
|
|
"papermill": {
|
|
"duration": 0.410347,
|
|
"end_time": "2026-03-24T02:22:45.078863+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:22:44.668516+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"outputs": [
|
|
{
|
|
"name": "stdout",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"6-Month Sharpe Autocorrelation:\n",
|
|
"shape: (3, 3)\n",
|
|
"┌─────┬──────────┬──────────┐\n",
|
|
"│ lag ┆ mean_acf ┆ std_acf │\n",
|
|
"│ --- ┆ --- ┆ --- │\n",
|
|
"│ i64 ┆ f64 ┆ f64 │\n",
|
|
"╞═════╪══════════╪══════════╡\n",
|
|
"│ 1 ┆ 0.988964 ┆ 0.002837 │\n",
|
|
"│ 5 ┆ 0.946673 ┆ 0.013299 │\n",
|
|
"│ 21 ┆ 0.794334 ┆ 0.047909 │\n",
|
|
"└─────┴──────────┴──────────┘\n"
|
|
]
|
|
}
|
|
],
|
|
"source": [
|
|
"acf_results = []\n",
|
|
"sample_assets = assets_list\n",
|
|
"for asset in sample_assets:\n",
|
|
" scores = features.filter(pl.col(\"symbol\") == asset)[\"sharpe_126d\"].drop_nulls().to_numpy()\n",
|
|
" for lag in [1, 5, 21]:\n",
|
|
" if len(scores) > lag + 10:\n",
|
|
" acf = np.corrcoef(scores[lag:], scores[:-lag])[0, 1]\n",
|
|
" acf_results.append({\"symbol\": asset, \"lag\": lag, \"acf\": acf})\n",
|
|
"\n",
|
|
"acf_df = pl.DataFrame(acf_results)\n",
|
|
"acf_summary = acf_df.group_by(\"lag\").agg(\n",
|
|
" pl.col(\"acf\").mean().alias(\"mean_acf\"),\n",
|
|
" pl.col(\"acf\").std().alias(\"std_acf\"),\n",
|
|
")\n",
|
|
"print(\"6-Month Sharpe Autocorrelation:\")\n",
|
|
"print(acf_summary.sort(\"lag\"))"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "19d65232",
|
|
"metadata": {
|
|
"papermill": {
|
|
"duration": 0.002161,
|
|
"end_time": "2026-03-24T02:22:45.084067+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:22:45.081906+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"source": [
|
|
"### 4.2 Rank Stability"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 15,
|
|
"id": "cd3f5a80",
|
|
"metadata": {
|
|
"execution": {
|
|
"iopub.execute_input": "2026-04-09T01:06:27.640456Z",
|
|
"iopub.status.busy": "2026-04-09T01:06:27.640361Z",
|
|
"iopub.status.idle": "2026-04-09T01:06:27.677955Z",
|
|
"shell.execute_reply": "2026-04-09T01:06:27.676778Z"
|
|
},
|
|
"papermill": {
|
|
"duration": 0.023201,
|
|
"end_time": "2026-03-24T02:22:45.109388+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:22:45.086187+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"outputs": [
|
|
{
|
|
"name": "stdout",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"Monthly rank persistence (correlation): 0.813\n"
|
|
]
|
|
}
|
|
],
|
|
"source": [
|
|
"monthly = (\n",
|
|
" features.with_columns(pl.col(\"timestamp\").dt.truncate(\"1mo\").alias(\"month\"))\n",
|
|
" .group_by([\"month\", \"symbol\"])\n",
|
|
" .agg(pl.col(\"sharpe_126d_rank\").last().alias(\"rank_end\"))\n",
|
|
" .sort([\"symbol\", \"month\"])\n",
|
|
" .with_columns(pl.col(\"rank_end\").shift(1).over(\"symbol\").alias(\"rank_prev\"))\n",
|
|
" .drop_nulls()\n",
|
|
")\n",
|
|
"\n",
|
|
"rank_corr = np.corrcoef(monthly[\"rank_prev\"].to_numpy(), monthly[\"rank_end\"].to_numpy())[0, 1]\n",
|
|
"print(f\"Monthly rank persistence (correlation): {rank_corr:.3f}\")"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "5d50196b",
|
|
"metadata": {
|
|
"papermill": {
|
|
"duration": 0.002282,
|
|
"end_time": "2026-03-24T02:22:45.114184+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:22:45.111902+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"source": [
|
|
"## 5. Save Features"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 16,
|
|
"id": "e8de2b2d",
|
|
"metadata": {
|
|
"execution": {
|
|
"iopub.execute_input": "2026-04-09T01:06:27.679378Z",
|
|
"iopub.status.busy": "2026-04-09T01:06:27.679227Z",
|
|
"iopub.status.idle": "2026-04-09T01:06:27.966340Z",
|
|
"shell.execute_reply": "2026-04-09T01:06:27.965775Z"
|
|
},
|
|
"lines_to_next_cell": 0,
|
|
"papermill": {
|
|
"duration": 0.169856,
|
|
"end_time": "2026-03-24T02:22:45.286188+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:22:45.116332+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"outputs": [
|
|
{
|
|
"name": "stdout",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"Saved 57 features to case_studies/etfs/features/financial.parquet\n",
|
|
" 396,186 rows, 99 assets\n"
|
|
]
|
|
}
|
|
],
|
|
"source": [
|
|
"FEATURES_DIR.mkdir(parents=True, exist_ok=True)\n",
|
|
"features.write_parquet(FEATURES_DIR / \"financial.parquet\")\n",
|
|
"print(f\"Saved {len(feature_cols)} features to {FEATURES_DIR / 'financial.parquet'}\")\n",
|
|
"print(f\" {len(features):,} rows, {features['symbol'].n_unique()} assets\")"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "dd48b04d",
|
|
"metadata": {
|
|
"papermill": {
|
|
"duration": 0.002936,
|
|
"end_time": "2026-03-24T02:22:45.291623+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:22:45.288687+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"source": [
|
|
"**Note on frequency**: The feature matrix is saved at daily frequency\n",
|
|
"(all trading days), while the decision cadence is monthly month-end.\n",
|
|
"This means ~21 rows per decision point. Features are saved daily for\n",
|
|
"flexibility (enabling weekly variant analysis), but downstream modeling\n",
|
|
"(Ch11+) should sample at the decision cadence. The IC evaluation below\n",
|
|
"uses all daily cross-sections with HAC adjustment for the resulting\n",
|
|
"autocorrelation in overlapping 21-day labels."
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "86498def",
|
|
"metadata": {
|
|
"papermill": {
|
|
"duration": 0.002074,
|
|
"end_time": "2026-03-24T02:22:45.295925+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:22:45.293851+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"source": [
|
|
"## 6. Feature Evaluation\n",
|
|
"\n",
|
|
"We evaluate each feature's predictive power using cross-sectional Spearman IC\n",
|
|
"against 21-day forward returns. Two critical adjustments:\n",
|
|
"\n",
|
|
"- **HAC standard errors** (Newey-West): Overlapping 21-day returns create\n",
|
|
" autocorrelated IC time series, inflating naive t-statistics\n",
|
|
"- **BH-FDR correction**: With 57 simultaneous tests, naive p-values overstate\n",
|
|
" significance; we control the false discovery rate at 5%\n",
|
|
"\n",
|
|
"The cross-sectional IC on date $t$ is:\n",
|
|
"\n",
|
|
"$$IC_t = \\rho_S\\bigl(f_{t,i},\\; r_{t \\to t+21, i}\\bigr)$$\n",
|
|
"\n",
|
|
"where $f_{t,i}$ is the feature value for asset $i$ on date $t$.\n",
|
|
"\n",
|
|
"**Caveat**: This is a preliminary evaluation of financial features only.\n",
|
|
"The authoritative evaluation — covering both financial and temporal features\n",
|
|
"with consistent methodology — is in [`05_evaluation`](05_evaluation.ipynb). Results may differ\n",
|
|
"due to different feature sets and evaluation scope."
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 17,
|
|
"id": "024e6c65",
|
|
"metadata": {
|
|
"execution": {
|
|
"iopub.execute_input": "2026-04-09T01:06:27.967633Z",
|
|
"iopub.status.busy": "2026-04-09T01:06:27.967501Z",
|
|
"iopub.status.idle": "2026-04-09T01:06:28.071903Z",
|
|
"shell.execute_reply": "2026-04-09T01:06:28.067105Z"
|
|
},
|
|
"papermill": {
|
|
"duration": 0.055015,
|
|
"end_time": "2026-03-24T02:22:45.353061+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:22:45.298046+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"outputs": [
|
|
{
|
|
"name": "stdout",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"Evaluation set: 394,233 rows (4,759 dates x 99 assets)\n"
|
|
]
|
|
}
|
|
],
|
|
"source": [
|
|
"# Load 21-day forward return labels\n",
|
|
"labels = pl.read_parquet(CASE_DIR / \"labels\" / \"fwd_ret_21d.parquet\")\n",
|
|
"label_col = \"fwd_ret_21d\"\n",
|
|
"\n",
|
|
"eval_df = features.join(\n",
|
|
" labels.select([\"timestamp\", \"symbol\", label_col]), on=[\"timestamp\", \"symbol\"], how=\"inner\"\n",
|
|
")\n",
|
|
"print(\n",
|
|
" f\"Evaluation set: {len(eval_df):,} rows \"\n",
|
|
" f\"({eval_df['timestamp'].n_unique():,} dates x {eval_df['symbol'].n_unique()} assets)\"\n",
|
|
")"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "e121481f",
|
|
"metadata": {
|
|
"papermill": {
|
|
"duration": 0.00223,
|
|
"end_time": "2026-03-24T02:22:45.357902+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:22:45.355672+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"source": [
|
|
"### Cross-Sectional IC with HAC Adjustment"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 18,
|
|
"id": "ff45e482",
|
|
"metadata": {
|
|
"execution": {
|
|
"iopub.execute_input": "2026-04-09T01:06:28.079516Z",
|
|
"iopub.status.busy": "2026-04-09T01:06:28.079137Z",
|
|
"iopub.status.idle": "2026-04-09T01:07:51.740955Z",
|
|
"shell.execute_reply": "2026-04-09T01:07:51.740356Z"
|
|
},
|
|
"papermill": {
|
|
"duration": 57.733832,
|
|
"end_time": "2026-03-24T02:23:43.093871+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:22:45.360039+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"outputs": [
|
|
{
|
|
"name": "stdout",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"Computed IC stats for 53 features (4,759 dates)\n"
|
|
]
|
|
}
|
|
],
|
|
"source": [
|
|
"# Compute cross-sectional Spearman IC per feature per date\n",
|
|
"dates = eval_df[\"timestamp\"].unique().sort().to_list()\n",
|
|
"n_dates = len(dates)\n",
|
|
"ic_matrix = np.full((n_dates, len(feature_cols)), np.nan)\n",
|
|
"\n",
|
|
"for i, date_val in enumerate(dates):\n",
|
|
" group = eval_df.filter(pl.col(\"timestamp\") == date_val)\n",
|
|
" y = np.asarray(group[label_col].to_numpy(), dtype=float)\n",
|
|
" valid_y = ~np.isnan(y)\n",
|
|
" if valid_y.sum() < 10:\n",
|
|
" continue\n",
|
|
" for j, feat in enumerate(feature_cols):\n",
|
|
" if group[feat].dtype not in NUMERIC_DTYPES:\n",
|
|
" continue\n",
|
|
" x = np.asarray(group[feat].to_numpy(), dtype=float)\n",
|
|
" valid = valid_y & ~np.isnan(x)\n",
|
|
" if valid.sum() >= 5:\n",
|
|
" rho, _ = spearmanr(x[valid], y[valid])\n",
|
|
" ic_matrix[i, j] = rho\n",
|
|
"\n",
|
|
"# Build per-feature HAC statistics\n",
|
|
"ic_stats = {}\n",
|
|
"for j, feat in enumerate(feature_cols):\n",
|
|
" ic_series = ic_matrix[:, j]\n",
|
|
" valid_ic = ic_series[~np.isnan(ic_series)]\n",
|
|
" if len(valid_ic) < 30:\n",
|
|
" continue\n",
|
|
" stats = compute_ic_hac_stats(pl.DataFrame({\"ic\": valid_ic}), ic_col=\"ic\")\n",
|
|
" ic_stats[feat] = stats\n",
|
|
"\n",
|
|
"print(f\"Computed IC stats for {len(ic_stats)} features ({n_dates:,} dates)\")"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "55a6ee75",
|
|
"metadata": {
|
|
"papermill": {
|
|
"duration": 0.002255,
|
|
"end_time": "2026-03-24T02:23:43.107112+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:23:43.104857+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"source": [
|
|
"### BH-FDR Multiple Testing Correction"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 19,
|
|
"id": "4d0c906b",
|
|
"metadata": {
|
|
"execution": {
|
|
"iopub.execute_input": "2026-04-09T01:07:51.741929Z",
|
|
"iopub.status.busy": "2026-04-09T01:07:51.741849Z",
|
|
"iopub.status.idle": "2026-04-09T01:07:51.745881Z",
|
|
"shell.execute_reply": "2026-04-09T01:07:51.745508Z"
|
|
},
|
|
"papermill": {
|
|
"duration": 0.007766,
|
|
"end_time": "2026-03-24T02:23:43.117012+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:23:43.109246+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"outputs": [
|
|
{
|
|
"name": "stdout",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"\n",
|
|
"Feature Evaluation Summary\n",
|
|
" Features tested: 53\n",
|
|
" Naive significant (p<0.05): 16\n",
|
|
" FDR-significant (q<0.05): 13\n",
|
|
" Inflation factor: 1.2x\n",
|
|
"\n",
|
|
"Top 10 by IC:\n",
|
|
"shape: (10, 7)\n",
|
|
"┌──────────────────┬──────────┬───────────┬─────────────┬───────────┬────────────┬─────────────────┐\n",
|
|
"│ feature ┆ ic_mean ┆ hac_tstat ┆ naive_tstat ┆ p_value ┆ adjusted_p ┆ significant_fdr │\n",
|
|
"│ --- ┆ --- ┆ --- ┆ --- ┆ --- ┆ --- ┆ 05 │\n",
|
|
"│ str ┆ f64 ┆ f64 ┆ f64 ┆ f64 ┆ f64 ┆ --- │\n",
|
|
"│ ┆ ┆ ┆ ┆ ┆ ┆ bool │\n",
|
|
"╞══════════════════╪══════════╪═══════════╪═════════════╪═══════════╪════════════╪═════════════════╡\n",
|
|
"│ dist_52w_low ┆ 0.076355 ┆ 5.834005 ┆ 16.152049 ┆ 5.7695e-9 ┆ 3.0578e-7 ┆ true │\n",
|
|
"│ vol_252d ┆ 0.058948 ┆ 3.67656 ┆ 10.401608 ┆ 0.000239 ┆ 0.004222 ┆ true │\n",
|
|
"│ natr_14 ┆ 0.05722 ┆ 3.704387 ┆ 10.388454 ┆ 0.000214 ┆ 0.004222 ┆ true │\n",
|
|
"│ vol_126d ┆ 0.056434 ┆ 3.555592 ┆ 10.029753 ┆ 0.000381 ┆ 0.005046 ┆ true │\n",
|
|
"│ vol_63d ┆ 0.05091 ┆ 3.239456 ┆ 9.110108 ┆ 0.001206 ┆ 0.008736 ┆ true │\n",
|
|
"│ vol_63d_rank ┆ 0.05091 ┆ 3.239456 ┆ 9.110108 ┆ 0.001206 ┆ 0.008736 ┆ true │\n",
|
|
"│ vol_21d ┆ 0.050691 ┆ 3.302222 ┆ 9.232101 ┆ 0.000966 ┆ 0.008736 ┆ true │\n",
|
|
"│ skip_recent_12_1 ┆ 0.041785 ┆ 2.919424 ┆ 8.21859 ┆ 0.003523 ┆ 0.017045 ┆ true │\n",
|
|
"│ ret_189d ┆ 0.041107 ┆ 2.918157 ┆ 8.220979 ┆ 0.003538 ┆ 0.017045 ┆ true │\n",
|
|
"│ ret_252d ┆ 0.040336 ┆ 2.838828 ┆ 7.997445 ┆ 0.004547 ┆ 0.020083 ┆ true │\n",
|
|
"└──────────────────┴──────────┴───────────┴─────────────┴───────────┴────────────┴─────────────────┘\n"
|
|
]
|
|
}
|
|
],
|
|
"source": [
|
|
"feat_names = list(ic_stats.keys())\n",
|
|
"p_values = [ic_stats[f][\"p_value\"] for f in feat_names]\n",
|
|
"fdr_result = benjamini_hochberg_fdr(p_values, alpha=0.05, return_details=True)\n",
|
|
"\n",
|
|
"eval_summary = pl.DataFrame(\n",
|
|
" {\n",
|
|
" \"feature\": feat_names,\n",
|
|
" \"ic_mean\": [ic_stats[f][\"mean_ic\"] for f in feat_names],\n",
|
|
" \"hac_tstat\": [ic_stats[f][\"t_stat\"] for f in feat_names],\n",
|
|
" \"naive_tstat\": [ic_stats[f][\"naive_t_stat\"] for f in feat_names],\n",
|
|
" \"p_value\": p_values,\n",
|
|
" \"adjusted_p\": fdr_result[\"adjusted_p_values\"].tolist(),\n",
|
|
" \"significant_fdr05\": fdr_result[\"rejected\"].tolist(),\n",
|
|
" }\n",
|
|
").sort(\"ic_mean\", descending=True)\n",
|
|
"\n",
|
|
"n_naive_sig = sum(1 for p in p_values if p < 0.05)\n",
|
|
"n_fdr_sig = int(fdr_result[\"n_rejected\"])\n",
|
|
"inflation = n_naive_sig / max(n_fdr_sig, 1)\n",
|
|
"\n",
|
|
"print(\"\\nFeature Evaluation Summary\")\n",
|
|
"print(f\" Features tested: {len(feat_names)}\")\n",
|
|
"print(f\" Naive significant (p<0.05): {n_naive_sig}\")\n",
|
|
"print(f\" FDR-significant (q<0.05): {n_fdr_sig}\")\n",
|
|
"print(f\" Inflation factor: {inflation:.1f}x\")\n",
|
|
"print(\"\\nTop 10 by IC:\")\n",
|
|
"print(eval_summary.head(10))"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "9788628f",
|
|
"metadata": {
|
|
"papermill": {
|
|
"duration": 0.003556,
|
|
"end_time": "2026-03-24T02:23:43.124286+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:23:43.120730+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"source": [
|
|
"**Note on date-level features**: Features like `regime_binary`,\n",
|
|
"`yield_curve_slope`, `corr_spy_tlt_63d`, and `choppiness_63d` are identical\n",
|
|
"across all symbols on a given date. Their cross-sectional IC is zero by\n",
|
|
"construction because they cannot differentiate between assets. These features\n",
|
|
"add value through interactions with cross-sectional features (e.g., conditional\n",
|
|
"momentum) rather than standalone ranking ability."
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "64458903",
|
|
"metadata": {
|
|
"papermill": {
|
|
"duration": 0.002123,
|
|
"end_time": "2026-03-24T02:23:43.129048+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:23:43.126925+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"source": [
|
|
"### IC Bar Chart"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 20,
|
|
"id": "764e5a56",
|
|
"metadata": {
|
|
"execution": {
|
|
"iopub.execute_input": "2026-04-09T01:07:51.746786Z",
|
|
"iopub.status.busy": "2026-04-09T01:07:51.746712Z",
|
|
"iopub.status.idle": "2026-04-09T01:07:52.724425Z",
|
|
"shell.execute_reply": "2026-04-09T01:07:52.723999Z"
|
|
},
|
|
"papermill": {
|
|
"duration": 1.062994,
|
|
"end_time": "2026-03-24T02:23:44.194239+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:23:43.131245+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"outputs": [
|
|
{
|
|
"data": {
|
|
"application/vnd.plotly.v1+json": {
|
|
"config": {
|
|
"plotlyServerURL": "https://plot.ly"
|
|
},
|
|
"data": [
|
|
{
|
|
"marker": {
|
|
"color": [
|
|
"#2ecc71",
|
|
"#2ecc71",
|
|
"#2ecc71",
|
|
"#2ecc71",
|
|
"#2ecc71",
|
|
"#2ecc71",
|
|
"#2ecc71",
|
|
"#2ecc71",
|
|
"#2ecc71",
|
|
"#2ecc71",
|
|
"#2ecc71",
|
|
"#95a5a6",
|
|
"#2ecc71",
|
|
"#95a5a6",
|
|
"#95a5a6",
|
|
"#95a5a6",
|
|
"#95a5a6",
|
|
"#95a5a6",
|
|
"#95a5a6",
|
|
"#95a5a6"
|
|
]
|
|
},
|
|
"text": [
|
|
"0.076",
|
|
"0.059",
|
|
"0.057",
|
|
"0.056",
|
|
"0.051",
|
|
"0.051",
|
|
"0.051",
|
|
"-0.043",
|
|
"0.042",
|
|
"0.041",
|
|
"0.040",
|
|
"0.034",
|
|
"0.029",
|
|
"-0.026",
|
|
"0.024",
|
|
"0.023",
|
|
"0.023",
|
|
"-0.020",
|
|
"0.020",
|
|
"-0.020"
|
|
],
|
|
"textposition": "outside",
|
|
"type": "bar",
|
|
"x": [
|
|
"dist_52w_low",
|
|
"vol_252d",
|
|
"natr_14",
|
|
"vol_126d",
|
|
"vol_63d",
|
|
"vol_63d_rank",
|
|
"vol_21d",
|
|
"mom_accel_long",
|
|
"skip_recent_12_1",
|
|
"ret_189d",
|
|
"ret_252d",
|
|
"sma_ratio_200",
|
|
"pct_positive_63d",
|
|
"mom_accel_short",
|
|
"skip_recent_6_1",
|
|
"ret_126d",
|
|
"ret_126d_rank",
|
|
"max_dd_63d",
|
|
"ret_63d",
|
|
"mom_accel_medium"
|
|
],
|
|
"y": [
|
|
0.07635483496346775,
|
|
0.05894811022065204,
|
|
0.05721956547277099,
|
|
0.056434302982184055,
|
|
0.05090995251982655,
|
|
0.05090995251982655,
|
|
0.050691280996018585,
|
|
-0.04325755427324245,
|
|
0.04178482635486817,
|
|
0.04110730011268391,
|
|
0.040335762919327035,
|
|
0.033548199627978344,
|
|
0.029383844733022125,
|
|
-0.02609920732671317,
|
|
0.02408747298706548,
|
|
0.02338393916006296,
|
|
0.02338393916006296,
|
|
-0.020294611560483112,
|
|
0.02014832251655958,
|
|
-0.020024239111810494
|
|
]
|
|
}
|
|
],
|
|
"layout": {
|
|
"height": 500,
|
|
"template": {
|
|
"data": {
|
|
"bar": [
|
|
{
|
|
"error_x": {
|
|
"color": "#2a3f5f"
|
|
},
|
|
"error_y": {
|
|
"color": "#2a3f5f"
|
|
},
|
|
"marker": {
|
|
"line": {
|
|
"color": "#E5ECF6",
|
|
"width": 0.5
|
|
},
|
|
"pattern": {
|
|
"fillmode": "overlay",
|
|
"size": 10,
|
|
"solidity": 0.2
|
|
}
|
|
},
|
|
"type": "bar"
|
|
}
|
|
],
|
|
"barpolar": [
|
|
{
|
|
"marker": {
|
|
"line": {
|
|
"color": "#E5ECF6",
|
|
"width": 0.5
|
|
},
|
|
"pattern": {
|
|
"fillmode": "overlay",
|
|
"size": 10,
|
|
"solidity": 0.2
|
|
}
|
|
},
|
|
"type": "barpolar"
|
|
}
|
|
],
|
|
"carpet": [
|
|
{
|
|
"aaxis": {
|
|
"endlinecolor": "#2a3f5f",
|
|
"gridcolor": "white",
|
|
"linecolor": "white",
|
|
"minorgridcolor": "white",
|
|
"startlinecolor": "#2a3f5f"
|
|
},
|
|
"baxis": {
|
|
"endlinecolor": "#2a3f5f",
|
|
"gridcolor": "white",
|
|
"linecolor": "white",
|
|
"minorgridcolor": "white",
|
|
"startlinecolor": "#2a3f5f"
|
|
},
|
|
"type": "carpet"
|
|
}
|
|
],
|
|
"choropleth": [
|
|
{
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
},
|
|
"type": "choropleth"
|
|
}
|
|
],
|
|
"contour": [
|
|
{
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
},
|
|
"colorscale": [
|
|
[
|
|
0.0,
|
|
"#0d0887"
|
|
],
|
|
[
|
|
0.1111111111111111,
|
|
"#46039f"
|
|
],
|
|
[
|
|
0.2222222222222222,
|
|
"#7201a8"
|
|
],
|
|
[
|
|
0.3333333333333333,
|
|
"#9c179e"
|
|
],
|
|
[
|
|
0.4444444444444444,
|
|
"#bd3786"
|
|
],
|
|
[
|
|
0.5555555555555556,
|
|
"#d8576b"
|
|
],
|
|
[
|
|
0.6666666666666666,
|
|
"#ed7953"
|
|
],
|
|
[
|
|
0.7777777777777778,
|
|
"#fb9f3a"
|
|
],
|
|
[
|
|
0.8888888888888888,
|
|
"#fdca26"
|
|
],
|
|
[
|
|
1.0,
|
|
"#f0f921"
|
|
]
|
|
],
|
|
"type": "contour"
|
|
}
|
|
],
|
|
"contourcarpet": [
|
|
{
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
},
|
|
"type": "contourcarpet"
|
|
}
|
|
],
|
|
"heatmap": [
|
|
{
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
},
|
|
"colorscale": [
|
|
[
|
|
0.0,
|
|
"#0d0887"
|
|
],
|
|
[
|
|
0.1111111111111111,
|
|
"#46039f"
|
|
],
|
|
[
|
|
0.2222222222222222,
|
|
"#7201a8"
|
|
],
|
|
[
|
|
0.3333333333333333,
|
|
"#9c179e"
|
|
],
|
|
[
|
|
0.4444444444444444,
|
|
"#bd3786"
|
|
],
|
|
[
|
|
0.5555555555555556,
|
|
"#d8576b"
|
|
],
|
|
[
|
|
0.6666666666666666,
|
|
"#ed7953"
|
|
],
|
|
[
|
|
0.7777777777777778,
|
|
"#fb9f3a"
|
|
],
|
|
[
|
|
0.8888888888888888,
|
|
"#fdca26"
|
|
],
|
|
[
|
|
1.0,
|
|
"#f0f921"
|
|
]
|
|
],
|
|
"type": "heatmap"
|
|
}
|
|
],
|
|
"histogram": [
|
|
{
|
|
"marker": {
|
|
"pattern": {
|
|
"fillmode": "overlay",
|
|
"size": 10,
|
|
"solidity": 0.2
|
|
}
|
|
},
|
|
"type": "histogram"
|
|
}
|
|
],
|
|
"histogram2d": [
|
|
{
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
},
|
|
"colorscale": [
|
|
[
|
|
0.0,
|
|
"#0d0887"
|
|
],
|
|
[
|
|
0.1111111111111111,
|
|
"#46039f"
|
|
],
|
|
[
|
|
0.2222222222222222,
|
|
"#7201a8"
|
|
],
|
|
[
|
|
0.3333333333333333,
|
|
"#9c179e"
|
|
],
|
|
[
|
|
0.4444444444444444,
|
|
"#bd3786"
|
|
],
|
|
[
|
|
0.5555555555555556,
|
|
"#d8576b"
|
|
],
|
|
[
|
|
0.6666666666666666,
|
|
"#ed7953"
|
|
],
|
|
[
|
|
0.7777777777777778,
|
|
"#fb9f3a"
|
|
],
|
|
[
|
|
0.8888888888888888,
|
|
"#fdca26"
|
|
],
|
|
[
|
|
1.0,
|
|
"#f0f921"
|
|
]
|
|
],
|
|
"type": "histogram2d"
|
|
}
|
|
],
|
|
"histogram2dcontour": [
|
|
{
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
},
|
|
"colorscale": [
|
|
[
|
|
0.0,
|
|
"#0d0887"
|
|
],
|
|
[
|
|
0.1111111111111111,
|
|
"#46039f"
|
|
],
|
|
[
|
|
0.2222222222222222,
|
|
"#7201a8"
|
|
],
|
|
[
|
|
0.3333333333333333,
|
|
"#9c179e"
|
|
],
|
|
[
|
|
0.4444444444444444,
|
|
"#bd3786"
|
|
],
|
|
[
|
|
0.5555555555555556,
|
|
"#d8576b"
|
|
],
|
|
[
|
|
0.6666666666666666,
|
|
"#ed7953"
|
|
],
|
|
[
|
|
0.7777777777777778,
|
|
"#fb9f3a"
|
|
],
|
|
[
|
|
0.8888888888888888,
|
|
"#fdca26"
|
|
],
|
|
[
|
|
1.0,
|
|
"#f0f921"
|
|
]
|
|
],
|
|
"type": "histogram2dcontour"
|
|
}
|
|
],
|
|
"mesh3d": [
|
|
{
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
},
|
|
"type": "mesh3d"
|
|
}
|
|
],
|
|
"parcoords": [
|
|
{
|
|
"line": {
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
}
|
|
},
|
|
"type": "parcoords"
|
|
}
|
|
],
|
|
"pie": [
|
|
{
|
|
"automargin": true,
|
|
"type": "pie"
|
|
}
|
|
],
|
|
"scatter": [
|
|
{
|
|
"fillpattern": {
|
|
"fillmode": "overlay",
|
|
"size": 10,
|
|
"solidity": 0.2
|
|
},
|
|
"type": "scatter"
|
|
}
|
|
],
|
|
"scatter3d": [
|
|
{
|
|
"line": {
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
}
|
|
},
|
|
"marker": {
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
}
|
|
},
|
|
"type": "scatter3d"
|
|
}
|
|
],
|
|
"scattercarpet": [
|
|
{
|
|
"marker": {
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
}
|
|
},
|
|
"type": "scattercarpet"
|
|
}
|
|
],
|
|
"scattergeo": [
|
|
{
|
|
"marker": {
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
}
|
|
},
|
|
"type": "scattergeo"
|
|
}
|
|
],
|
|
"scattergl": [
|
|
{
|
|
"marker": {
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
}
|
|
},
|
|
"type": "scattergl"
|
|
}
|
|
],
|
|
"scattermap": [
|
|
{
|
|
"marker": {
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
}
|
|
},
|
|
"type": "scattermap"
|
|
}
|
|
],
|
|
"scattermapbox": [
|
|
{
|
|
"marker": {
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
}
|
|
},
|
|
"type": "scattermapbox"
|
|
}
|
|
],
|
|
"scatterpolar": [
|
|
{
|
|
"marker": {
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
}
|
|
},
|
|
"type": "scatterpolar"
|
|
}
|
|
],
|
|
"scatterpolargl": [
|
|
{
|
|
"marker": {
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
}
|
|
},
|
|
"type": "scatterpolargl"
|
|
}
|
|
],
|
|
"scatterternary": [
|
|
{
|
|
"marker": {
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
}
|
|
},
|
|
"type": "scatterternary"
|
|
}
|
|
],
|
|
"surface": [
|
|
{
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
},
|
|
"colorscale": [
|
|
[
|
|
0.0,
|
|
"#0d0887"
|
|
],
|
|
[
|
|
0.1111111111111111,
|
|
"#46039f"
|
|
],
|
|
[
|
|
0.2222222222222222,
|
|
"#7201a8"
|
|
],
|
|
[
|
|
0.3333333333333333,
|
|
"#9c179e"
|
|
],
|
|
[
|
|
0.4444444444444444,
|
|
"#bd3786"
|
|
],
|
|
[
|
|
0.5555555555555556,
|
|
"#d8576b"
|
|
],
|
|
[
|
|
0.6666666666666666,
|
|
"#ed7953"
|
|
],
|
|
[
|
|
0.7777777777777778,
|
|
"#fb9f3a"
|
|
],
|
|
[
|
|
0.8888888888888888,
|
|
"#fdca26"
|
|
],
|
|
[
|
|
1.0,
|
|
"#f0f921"
|
|
]
|
|
],
|
|
"type": "surface"
|
|
}
|
|
],
|
|
"table": [
|
|
{
|
|
"cells": {
|
|
"fill": {
|
|
"color": "#EBF0F8"
|
|
},
|
|
"line": {
|
|
"color": "white"
|
|
}
|
|
},
|
|
"header": {
|
|
"fill": {
|
|
"color": "#C8D4E3"
|
|
},
|
|
"line": {
|
|
"color": "white"
|
|
}
|
|
},
|
|
"type": "table"
|
|
}
|
|
]
|
|
},
|
|
"layout": {
|
|
"annotationdefaults": {
|
|
"arrowcolor": "#2a3f5f",
|
|
"arrowhead": 0,
|
|
"arrowwidth": 1
|
|
},
|
|
"autotypenumbers": "strict",
|
|
"coloraxis": {
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
}
|
|
},
|
|
"colorscale": {
|
|
"diverging": [
|
|
[
|
|
0,
|
|
"#8e0152"
|
|
],
|
|
[
|
|
0.1,
|
|
"#c51b7d"
|
|
],
|
|
[
|
|
0.2,
|
|
"#de77ae"
|
|
],
|
|
[
|
|
0.3,
|
|
"#f1b6da"
|
|
],
|
|
[
|
|
0.4,
|
|
"#fde0ef"
|
|
],
|
|
[
|
|
0.5,
|
|
"#f7f7f7"
|
|
],
|
|
[
|
|
0.6,
|
|
"#e6f5d0"
|
|
],
|
|
[
|
|
0.7,
|
|
"#b8e186"
|
|
],
|
|
[
|
|
0.8,
|
|
"#7fbc41"
|
|
],
|
|
[
|
|
0.9,
|
|
"#4d9221"
|
|
],
|
|
[
|
|
1,
|
|
"#276419"
|
|
]
|
|
],
|
|
"sequential": [
|
|
[
|
|
0.0,
|
|
"#0d0887"
|
|
],
|
|
[
|
|
0.1111111111111111,
|
|
"#46039f"
|
|
],
|
|
[
|
|
0.2222222222222222,
|
|
"#7201a8"
|
|
],
|
|
[
|
|
0.3333333333333333,
|
|
"#9c179e"
|
|
],
|
|
[
|
|
0.4444444444444444,
|
|
"#bd3786"
|
|
],
|
|
[
|
|
0.5555555555555556,
|
|
"#d8576b"
|
|
],
|
|
[
|
|
0.6666666666666666,
|
|
"#ed7953"
|
|
],
|
|
[
|
|
0.7777777777777778,
|
|
"#fb9f3a"
|
|
],
|
|
[
|
|
0.8888888888888888,
|
|
"#fdca26"
|
|
],
|
|
[
|
|
1.0,
|
|
"#f0f921"
|
|
]
|
|
],
|
|
"sequentialminus": [
|
|
[
|
|
0.0,
|
|
"#0d0887"
|
|
],
|
|
[
|
|
0.1111111111111111,
|
|
"#46039f"
|
|
],
|
|
[
|
|
0.2222222222222222,
|
|
"#7201a8"
|
|
],
|
|
[
|
|
0.3333333333333333,
|
|
"#9c179e"
|
|
],
|
|
[
|
|
0.4444444444444444,
|
|
"#bd3786"
|
|
],
|
|
[
|
|
0.5555555555555556,
|
|
"#d8576b"
|
|
],
|
|
[
|
|
0.6666666666666666,
|
|
"#ed7953"
|
|
],
|
|
[
|
|
0.7777777777777778,
|
|
"#fb9f3a"
|
|
],
|
|
[
|
|
0.8888888888888888,
|
|
"#fdca26"
|
|
],
|
|
[
|
|
1.0,
|
|
"#f0f921"
|
|
]
|
|
]
|
|
},
|
|
"colorway": [
|
|
"#636efa",
|
|
"#EF553B",
|
|
"#00cc96",
|
|
"#ab63fa",
|
|
"#FFA15A",
|
|
"#19d3f3",
|
|
"#FF6692",
|
|
"#B6E880",
|
|
"#FF97FF",
|
|
"#FECB52"
|
|
],
|
|
"font": {
|
|
"color": "#2a3f5f"
|
|
},
|
|
"geo": {
|
|
"bgcolor": "white",
|
|
"lakecolor": "white",
|
|
"landcolor": "#E5ECF6",
|
|
"showlakes": true,
|
|
"showland": true,
|
|
"subunitcolor": "white"
|
|
},
|
|
"hoverlabel": {
|
|
"align": "left"
|
|
},
|
|
"hovermode": "closest",
|
|
"mapbox": {
|
|
"style": "light"
|
|
},
|
|
"paper_bgcolor": "white",
|
|
"plot_bgcolor": "#E5ECF6",
|
|
"polar": {
|
|
"angularaxis": {
|
|
"gridcolor": "white",
|
|
"linecolor": "white",
|
|
"ticks": ""
|
|
},
|
|
"bgcolor": "#E5ECF6",
|
|
"radialaxis": {
|
|
"gridcolor": "white",
|
|
"linecolor": "white",
|
|
"ticks": ""
|
|
}
|
|
},
|
|
"scene": {
|
|
"xaxis": {
|
|
"backgroundcolor": "#E5ECF6",
|
|
"gridcolor": "white",
|
|
"gridwidth": 2,
|
|
"linecolor": "white",
|
|
"showbackground": true,
|
|
"ticks": "",
|
|
"zerolinecolor": "white"
|
|
},
|
|
"yaxis": {
|
|
"backgroundcolor": "#E5ECF6",
|
|
"gridcolor": "white",
|
|
"gridwidth": 2,
|
|
"linecolor": "white",
|
|
"showbackground": true,
|
|
"ticks": "",
|
|
"zerolinecolor": "white"
|
|
},
|
|
"zaxis": {
|
|
"backgroundcolor": "#E5ECF6",
|
|
"gridcolor": "white",
|
|
"gridwidth": 2,
|
|
"linecolor": "white",
|
|
"showbackground": true,
|
|
"ticks": "",
|
|
"zerolinecolor": "white"
|
|
}
|
|
},
|
|
"shapedefaults": {
|
|
"line": {
|
|
"color": "#2a3f5f"
|
|
}
|
|
},
|
|
"ternary": {
|
|
"aaxis": {
|
|
"gridcolor": "white",
|
|
"linecolor": "white",
|
|
"ticks": ""
|
|
},
|
|
"baxis": {
|
|
"gridcolor": "white",
|
|
"linecolor": "white",
|
|
"ticks": ""
|
|
},
|
|
"bgcolor": "#E5ECF6",
|
|
"caxis": {
|
|
"gridcolor": "white",
|
|
"linecolor": "white",
|
|
"ticks": ""
|
|
}
|
|
},
|
|
"title": {
|
|
"x": 0.05
|
|
},
|
|
"xaxis": {
|
|
"automargin": true,
|
|
"gridcolor": "white",
|
|
"linecolor": "white",
|
|
"ticks": "",
|
|
"title": {
|
|
"standoff": 15
|
|
},
|
|
"zerolinecolor": "white",
|
|
"zerolinewidth": 2
|
|
},
|
|
"yaxis": {
|
|
"automargin": true,
|
|
"gridcolor": "white",
|
|
"linecolor": "white",
|
|
"ticks": "",
|
|
"title": {
|
|
"standoff": 15
|
|
},
|
|
"zerolinecolor": "white",
|
|
"zerolinewidth": 2
|
|
}
|
|
}
|
|
},
|
|
"title": {
|
|
"text": "Top 20 Features by |IC| (green = FDR-significant at 5%)"
|
|
},
|
|
"xaxis": {
|
|
"tickangle": 45,
|
|
"title": {
|
|
"text": "Feature"
|
|
}
|
|
},
|
|
"yaxis": {
|
|
"title": {
|
|
"text": "Mean IC"
|
|
}
|
|
}
|
|
}
|
|
},
|
|
"image/png": "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"
|
|
},
|
|
"metadata": {},
|
|
"output_type": "display_data"
|
|
}
|
|
],
|
|
"source": [
|
|
"top20 = eval_summary.sort(pl.col(\"ic_mean\").abs(), descending=True).head(20)\n",
|
|
"colors = [\"#2ecc71\" if sig else \"#95a5a6\" for sig in top20[\"significant_fdr05\"].to_list()]\n",
|
|
"\n",
|
|
"fig_ic = go.Figure(\n",
|
|
" go.Bar(\n",
|
|
" x=top20[\"feature\"].to_list(),\n",
|
|
" y=top20[\"ic_mean\"].to_list(),\n",
|
|
" marker_color=colors,\n",
|
|
" text=[f\"{ic:.3f}\" for ic in top20[\"ic_mean\"].to_list()],\n",
|
|
" textposition=\"outside\",\n",
|
|
" )\n",
|
|
")\n",
|
|
"fig_ic.update_layout(\n",
|
|
" title=\"Top 20 Features by |IC| (green = FDR-significant at 5%)\",\n",
|
|
" xaxis_title=\"Feature\",\n",
|
|
" yaxis_title=\"Mean IC\",\n",
|
|
" xaxis_tickangle=45,\n",
|
|
" height=500,\n",
|
|
")\n",
|
|
"fig_ic.show()"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "56ca1018",
|
|
"metadata": {
|
|
"papermill": {
|
|
"duration": 0.002816,
|
|
"end_time": "2026-03-24T02:23:44.200110+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:23:44.197294+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"source": [
|
|
"### Naive vs HAC t-Statistics"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 21,
|
|
"id": "68c614c2",
|
|
"metadata": {
|
|
"execution": {
|
|
"iopub.execute_input": "2026-04-09T01:07:52.725340Z",
|
|
"iopub.status.busy": "2026-04-09T01:07:52.725243Z",
|
|
"iopub.status.idle": "2026-04-09T01:07:52.771087Z",
|
|
"shell.execute_reply": "2026-04-09T01:07:52.770651Z"
|
|
},
|
|
"papermill": {
|
|
"duration": 0.012423,
|
|
"end_time": "2026-03-24T02:23:44.215312+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:23:44.202889+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"outputs": [
|
|
{
|
|
"data": {
|
|
"application/vnd.plotly.v1+json": {
|
|
"config": {
|
|
"plotlyServerURL": "https://plot.ly"
|
|
},
|
|
"data": [
|
|
{
|
|
"hovertemplate": "%{text}<br>Naive t: %{x:.2f}<br>HAC t: %{y:.2f}<extra></extra>",
|
|
"marker": {
|
|
"color": [
|
|
"#2ecc71",
|
|
"#2ecc71",
|
|
"#2ecc71",
|
|
"#2ecc71",
|
|
"#2ecc71",
|
|
"#2ecc71",
|
|
"#2ecc71",
|
|
"#2ecc71",
|
|
"#2ecc71",
|
|
"#2ecc71",
|
|
"#95a5a6",
|
|
"#2ecc71",
|
|
"#95a5a6",
|
|
"#95a5a6",
|
|
"#95a5a6",
|
|
"#95a5a6",
|
|
"#95a5a6",
|
|
"#95a5a6",
|
|
"#95a5a6",
|
|
"#2ecc71",
|
|
"#95a5a6",
|
|
"#95a5a6",
|
|
"#95a5a6",
|
|
"#95a5a6",
|
|
"#95a5a6",
|
|
"#95a5a6",
|
|
"#95a5a6",
|
|
"#95a5a6",
|
|
"#95a5a6",
|
|
"#95a5a6",
|
|
"#95a5a6",
|
|
"#95a5a6",
|
|
"#95a5a6",
|
|
"#95a5a6",
|
|
"#95a5a6",
|
|
"#95a5a6",
|
|
"#95a5a6",
|
|
"#95a5a6",
|
|
"#95a5a6",
|
|
"#95a5a6",
|
|
"#95a5a6",
|
|
"#95a5a6",
|
|
"#95a5a6",
|
|
"#95a5a6",
|
|
"#95a5a6",
|
|
"#95a5a6",
|
|
"#95a5a6",
|
|
"#95a5a6",
|
|
"#95a5a6",
|
|
"#95a5a6",
|
|
"#95a5a6",
|
|
"#95a5a6",
|
|
"#2ecc71"
|
|
],
|
|
"size": 8
|
|
},
|
|
"mode": "markers",
|
|
"text": [
|
|
"dist_52w_low",
|
|
"vol_252d",
|
|
"natr_14",
|
|
"vol_126d",
|
|
"vol_63d",
|
|
"vol_63d_rank",
|
|
"vol_21d",
|
|
"skip_recent_12_1",
|
|
"ret_189d",
|
|
"ret_252d",
|
|
"sma_ratio_200",
|
|
"pct_positive_63d",
|
|
"skip_recent_6_1",
|
|
"ret_126d",
|
|
"ret_126d_rank",
|
|
"ret_63d",
|
|
"sharpe_189d",
|
|
"sharpe_252d",
|
|
"vol_ratio_63d",
|
|
"vol_ratio_21d",
|
|
"sharpe_63d",
|
|
"hurst_100",
|
|
"vol_ratio_short",
|
|
"ret_42d",
|
|
"macd_line",
|
|
"sma_ratio_50",
|
|
"sharpe_126d",
|
|
"sharpe_126d_rank",
|
|
"ema_ratio_26",
|
|
"chop_14",
|
|
"sharpe_42d",
|
|
"ret_21d",
|
|
"ret_10d",
|
|
"dist_52w_high",
|
|
"adx_14",
|
|
"rsi_14",
|
|
"stoch_k",
|
|
"sharpe_21d",
|
|
"rsi_7",
|
|
"aroon_diff",
|
|
"sharpe_10d",
|
|
"ret_5d",
|
|
"sharpe_5d",
|
|
"cci_14",
|
|
"cci_21",
|
|
"obv_zscore_63d",
|
|
"bb_pctb_20",
|
|
"vol_ratio_medium",
|
|
"max_dd_126d",
|
|
"mom_accel_medium",
|
|
"max_dd_63d",
|
|
"mom_accel_short",
|
|
"mom_accel_long"
|
|
],
|
|
"type": "scatter",
|
|
"x": [
|
|
16.152049300038176,
|
|
10.4016075507356,
|
|
10.388454162685514,
|
|
10.029752836301446,
|
|
9.110108471151076,
|
|
9.110108471151076,
|
|
9.232100519933594,
|
|
8.218590383587179,
|
|
8.220978593170363,
|
|
7.997444789955963,
|
|
6.887362192807514,
|
|
8.186887848035505,
|
|
4.880859086460087,
|
|
4.765151116013638,
|
|
4.765151116013638,
|
|
4.160837207788063,
|
|
3.4409608081391028,
|
|
2.667595782857911,
|
|
4.685548712968539,
|
|
4.503854002005327,
|
|
1.7023350863586446,
|
|
2.210259353934044,
|
|
1.8968433729235636,
|
|
1.0167630917702568,
|
|
0.912949865812349,
|
|
0.5764095782358744,
|
|
0.586641688037098,
|
|
0.586641688037098,
|
|
0.11602668864737901,
|
|
0.0014786410254454996,
|
|
-0.5696204916308858,
|
|
-0.5431032988927165,
|
|
-1.1830453071757785,
|
|
-1.2454355612086927,
|
|
-2.0217830995934047,
|
|
-1.8104989769452282,
|
|
-1.9609186885402357,
|
|
-2.0358756479022198,
|
|
-2.2470797325178724,
|
|
-2.475057762826856,
|
|
-2.3656876386894536,
|
|
-2.123022458926743,
|
|
-2.627904830616295,
|
|
-2.851685560525629,
|
|
-3.369633808832531,
|
|
-4.4289020641172065,
|
|
-3.544083548045368,
|
|
-4.178142840935981,
|
|
-3.335857730998786,
|
|
-4.1065294803346735,
|
|
-4.253847606584472,
|
|
-5.390760069296785,
|
|
-9.031261408364609
|
|
],
|
|
"y": [
|
|
5.834005019092565,
|
|
3.6765596240559644,
|
|
3.7043872909673716,
|
|
3.555592461440577,
|
|
3.2394555703663532,
|
|
3.2394555703663532,
|
|
3.3022224684007857,
|
|
2.9194236402879574,
|
|
2.918156556048707,
|
|
2.8388275468469115,
|
|
2.474236613106108,
|
|
2.937766542448768,
|
|
1.747348851929744,
|
|
1.7115295450864112,
|
|
1.7115295450864112,
|
|
1.5081599494213997,
|
|
1.2178393307802715,
|
|
0.938059192142907,
|
|
2.448682615011774,
|
|
2.657100493202423,
|
|
0.6109556024524677,
|
|
0.805944302791836,
|
|
0.7248235863782979,
|
|
0.3805937615063145,
|
|
0.33606620274217774,
|
|
0.21843795338999042,
|
|
0.21008057130443764,
|
|
0.21008057130443764,
|
|
0.04674901467319806,
|
|
0.0006331314014872545,
|
|
-0.2094292533829732,
|
|
-0.21185456006750342,
|
|
-0.5012716633084335,
|
|
-0.45201844829752075,
|
|
-0.7285093028667483,
|
|
-0.7142616013391783,
|
|
-0.8648356441358066,
|
|
-0.7866126714130235,
|
|
-0.970219840517687,
|
|
-0.9496682299908759,
|
|
-0.9915025741432272,
|
|
-1.1138186484473513,
|
|
-1.355641518077316,
|
|
-1.2935594256603116,
|
|
-1.4152383696541506,
|
|
-1.6839853594872638,
|
|
-1.5068102705734643,
|
|
-1.498681127326364,
|
|
-1.2265919383787087,
|
|
-1.4822663322878429,
|
|
-1.5979827251125371,
|
|
-1.9605131891481573,
|
|
-3.2137996939842584
|
|
]
|
|
},
|
|
{
|
|
"line": {
|
|
"color": "gray",
|
|
"dash": "dash"
|
|
},
|
|
"mode": "lines",
|
|
"showlegend": false,
|
|
"type": "scatter",
|
|
"x": [
|
|
-9.531261408364609,
|
|
16.652049300038176
|
|
],
|
|
"y": [
|
|
-9.531261408364609,
|
|
16.652049300038176
|
|
]
|
|
}
|
|
],
|
|
"layout": {
|
|
"height": 500,
|
|
"template": {
|
|
"data": {
|
|
"bar": [
|
|
{
|
|
"error_x": {
|
|
"color": "#2a3f5f"
|
|
},
|
|
"error_y": {
|
|
"color": "#2a3f5f"
|
|
},
|
|
"marker": {
|
|
"line": {
|
|
"color": "#E5ECF6",
|
|
"width": 0.5
|
|
},
|
|
"pattern": {
|
|
"fillmode": "overlay",
|
|
"size": 10,
|
|
"solidity": 0.2
|
|
}
|
|
},
|
|
"type": "bar"
|
|
}
|
|
],
|
|
"barpolar": [
|
|
{
|
|
"marker": {
|
|
"line": {
|
|
"color": "#E5ECF6",
|
|
"width": 0.5
|
|
},
|
|
"pattern": {
|
|
"fillmode": "overlay",
|
|
"size": 10,
|
|
"solidity": 0.2
|
|
}
|
|
},
|
|
"type": "barpolar"
|
|
}
|
|
],
|
|
"carpet": [
|
|
{
|
|
"aaxis": {
|
|
"endlinecolor": "#2a3f5f",
|
|
"gridcolor": "white",
|
|
"linecolor": "white",
|
|
"minorgridcolor": "white",
|
|
"startlinecolor": "#2a3f5f"
|
|
},
|
|
"baxis": {
|
|
"endlinecolor": "#2a3f5f",
|
|
"gridcolor": "white",
|
|
"linecolor": "white",
|
|
"minorgridcolor": "white",
|
|
"startlinecolor": "#2a3f5f"
|
|
},
|
|
"type": "carpet"
|
|
}
|
|
],
|
|
"choropleth": [
|
|
{
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
},
|
|
"type": "choropleth"
|
|
}
|
|
],
|
|
"contour": [
|
|
{
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
},
|
|
"colorscale": [
|
|
[
|
|
0.0,
|
|
"#0d0887"
|
|
],
|
|
[
|
|
0.1111111111111111,
|
|
"#46039f"
|
|
],
|
|
[
|
|
0.2222222222222222,
|
|
"#7201a8"
|
|
],
|
|
[
|
|
0.3333333333333333,
|
|
"#9c179e"
|
|
],
|
|
[
|
|
0.4444444444444444,
|
|
"#bd3786"
|
|
],
|
|
[
|
|
0.5555555555555556,
|
|
"#d8576b"
|
|
],
|
|
[
|
|
0.6666666666666666,
|
|
"#ed7953"
|
|
],
|
|
[
|
|
0.7777777777777778,
|
|
"#fb9f3a"
|
|
],
|
|
[
|
|
0.8888888888888888,
|
|
"#fdca26"
|
|
],
|
|
[
|
|
1.0,
|
|
"#f0f921"
|
|
]
|
|
],
|
|
"type": "contour"
|
|
}
|
|
],
|
|
"contourcarpet": [
|
|
{
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
},
|
|
"type": "contourcarpet"
|
|
}
|
|
],
|
|
"heatmap": [
|
|
{
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
},
|
|
"colorscale": [
|
|
[
|
|
0.0,
|
|
"#0d0887"
|
|
],
|
|
[
|
|
0.1111111111111111,
|
|
"#46039f"
|
|
],
|
|
[
|
|
0.2222222222222222,
|
|
"#7201a8"
|
|
],
|
|
[
|
|
0.3333333333333333,
|
|
"#9c179e"
|
|
],
|
|
[
|
|
0.4444444444444444,
|
|
"#bd3786"
|
|
],
|
|
[
|
|
0.5555555555555556,
|
|
"#d8576b"
|
|
],
|
|
[
|
|
0.6666666666666666,
|
|
"#ed7953"
|
|
],
|
|
[
|
|
0.7777777777777778,
|
|
"#fb9f3a"
|
|
],
|
|
[
|
|
0.8888888888888888,
|
|
"#fdca26"
|
|
],
|
|
[
|
|
1.0,
|
|
"#f0f921"
|
|
]
|
|
],
|
|
"type": "heatmap"
|
|
}
|
|
],
|
|
"histogram": [
|
|
{
|
|
"marker": {
|
|
"pattern": {
|
|
"fillmode": "overlay",
|
|
"size": 10,
|
|
"solidity": 0.2
|
|
}
|
|
},
|
|
"type": "histogram"
|
|
}
|
|
],
|
|
"histogram2d": [
|
|
{
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
},
|
|
"colorscale": [
|
|
[
|
|
0.0,
|
|
"#0d0887"
|
|
],
|
|
[
|
|
0.1111111111111111,
|
|
"#46039f"
|
|
],
|
|
[
|
|
0.2222222222222222,
|
|
"#7201a8"
|
|
],
|
|
[
|
|
0.3333333333333333,
|
|
"#9c179e"
|
|
],
|
|
[
|
|
0.4444444444444444,
|
|
"#bd3786"
|
|
],
|
|
[
|
|
0.5555555555555556,
|
|
"#d8576b"
|
|
],
|
|
[
|
|
0.6666666666666666,
|
|
"#ed7953"
|
|
],
|
|
[
|
|
0.7777777777777778,
|
|
"#fb9f3a"
|
|
],
|
|
[
|
|
0.8888888888888888,
|
|
"#fdca26"
|
|
],
|
|
[
|
|
1.0,
|
|
"#f0f921"
|
|
]
|
|
],
|
|
"type": "histogram2d"
|
|
}
|
|
],
|
|
"histogram2dcontour": [
|
|
{
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
},
|
|
"colorscale": [
|
|
[
|
|
0.0,
|
|
"#0d0887"
|
|
],
|
|
[
|
|
0.1111111111111111,
|
|
"#46039f"
|
|
],
|
|
[
|
|
0.2222222222222222,
|
|
"#7201a8"
|
|
],
|
|
[
|
|
0.3333333333333333,
|
|
"#9c179e"
|
|
],
|
|
[
|
|
0.4444444444444444,
|
|
"#bd3786"
|
|
],
|
|
[
|
|
0.5555555555555556,
|
|
"#d8576b"
|
|
],
|
|
[
|
|
0.6666666666666666,
|
|
"#ed7953"
|
|
],
|
|
[
|
|
0.7777777777777778,
|
|
"#fb9f3a"
|
|
],
|
|
[
|
|
0.8888888888888888,
|
|
"#fdca26"
|
|
],
|
|
[
|
|
1.0,
|
|
"#f0f921"
|
|
]
|
|
],
|
|
"type": "histogram2dcontour"
|
|
}
|
|
],
|
|
"mesh3d": [
|
|
{
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
},
|
|
"type": "mesh3d"
|
|
}
|
|
],
|
|
"parcoords": [
|
|
{
|
|
"line": {
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
}
|
|
},
|
|
"type": "parcoords"
|
|
}
|
|
],
|
|
"pie": [
|
|
{
|
|
"automargin": true,
|
|
"type": "pie"
|
|
}
|
|
],
|
|
"scatter": [
|
|
{
|
|
"fillpattern": {
|
|
"fillmode": "overlay",
|
|
"size": 10,
|
|
"solidity": 0.2
|
|
},
|
|
"type": "scatter"
|
|
}
|
|
],
|
|
"scatter3d": [
|
|
{
|
|
"line": {
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
}
|
|
},
|
|
"marker": {
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
}
|
|
},
|
|
"type": "scatter3d"
|
|
}
|
|
],
|
|
"scattercarpet": [
|
|
{
|
|
"marker": {
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
}
|
|
},
|
|
"type": "scattercarpet"
|
|
}
|
|
],
|
|
"scattergeo": [
|
|
{
|
|
"marker": {
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
}
|
|
},
|
|
"type": "scattergeo"
|
|
}
|
|
],
|
|
"scattergl": [
|
|
{
|
|
"marker": {
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
}
|
|
},
|
|
"type": "scattergl"
|
|
}
|
|
],
|
|
"scattermap": [
|
|
{
|
|
"marker": {
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
}
|
|
},
|
|
"type": "scattermap"
|
|
}
|
|
],
|
|
"scattermapbox": [
|
|
{
|
|
"marker": {
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
}
|
|
},
|
|
"type": "scattermapbox"
|
|
}
|
|
],
|
|
"scatterpolar": [
|
|
{
|
|
"marker": {
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
}
|
|
},
|
|
"type": "scatterpolar"
|
|
}
|
|
],
|
|
"scatterpolargl": [
|
|
{
|
|
"marker": {
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
}
|
|
},
|
|
"type": "scatterpolargl"
|
|
}
|
|
],
|
|
"scatterternary": [
|
|
{
|
|
"marker": {
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
}
|
|
},
|
|
"type": "scatterternary"
|
|
}
|
|
],
|
|
"surface": [
|
|
{
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
},
|
|
"colorscale": [
|
|
[
|
|
0.0,
|
|
"#0d0887"
|
|
],
|
|
[
|
|
0.1111111111111111,
|
|
"#46039f"
|
|
],
|
|
[
|
|
0.2222222222222222,
|
|
"#7201a8"
|
|
],
|
|
[
|
|
0.3333333333333333,
|
|
"#9c179e"
|
|
],
|
|
[
|
|
0.4444444444444444,
|
|
"#bd3786"
|
|
],
|
|
[
|
|
0.5555555555555556,
|
|
"#d8576b"
|
|
],
|
|
[
|
|
0.6666666666666666,
|
|
"#ed7953"
|
|
],
|
|
[
|
|
0.7777777777777778,
|
|
"#fb9f3a"
|
|
],
|
|
[
|
|
0.8888888888888888,
|
|
"#fdca26"
|
|
],
|
|
[
|
|
1.0,
|
|
"#f0f921"
|
|
]
|
|
],
|
|
"type": "surface"
|
|
}
|
|
],
|
|
"table": [
|
|
{
|
|
"cells": {
|
|
"fill": {
|
|
"color": "#EBF0F8"
|
|
},
|
|
"line": {
|
|
"color": "white"
|
|
}
|
|
},
|
|
"header": {
|
|
"fill": {
|
|
"color": "#C8D4E3"
|
|
},
|
|
"line": {
|
|
"color": "white"
|
|
}
|
|
},
|
|
"type": "table"
|
|
}
|
|
]
|
|
},
|
|
"layout": {
|
|
"annotationdefaults": {
|
|
"arrowcolor": "#2a3f5f",
|
|
"arrowhead": 0,
|
|
"arrowwidth": 1
|
|
},
|
|
"autotypenumbers": "strict",
|
|
"coloraxis": {
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
}
|
|
},
|
|
"colorscale": {
|
|
"diverging": [
|
|
[
|
|
0,
|
|
"#8e0152"
|
|
],
|
|
[
|
|
0.1,
|
|
"#c51b7d"
|
|
],
|
|
[
|
|
0.2,
|
|
"#de77ae"
|
|
],
|
|
[
|
|
0.3,
|
|
"#f1b6da"
|
|
],
|
|
[
|
|
0.4,
|
|
"#fde0ef"
|
|
],
|
|
[
|
|
0.5,
|
|
"#f7f7f7"
|
|
],
|
|
[
|
|
0.6,
|
|
"#e6f5d0"
|
|
],
|
|
[
|
|
0.7,
|
|
"#b8e186"
|
|
],
|
|
[
|
|
0.8,
|
|
"#7fbc41"
|
|
],
|
|
[
|
|
0.9,
|
|
"#4d9221"
|
|
],
|
|
[
|
|
1,
|
|
"#276419"
|
|
]
|
|
],
|
|
"sequential": [
|
|
[
|
|
0.0,
|
|
"#0d0887"
|
|
],
|
|
[
|
|
0.1111111111111111,
|
|
"#46039f"
|
|
],
|
|
[
|
|
0.2222222222222222,
|
|
"#7201a8"
|
|
],
|
|
[
|
|
0.3333333333333333,
|
|
"#9c179e"
|
|
],
|
|
[
|
|
0.4444444444444444,
|
|
"#bd3786"
|
|
],
|
|
[
|
|
0.5555555555555556,
|
|
"#d8576b"
|
|
],
|
|
[
|
|
0.6666666666666666,
|
|
"#ed7953"
|
|
],
|
|
[
|
|
0.7777777777777778,
|
|
"#fb9f3a"
|
|
],
|
|
[
|
|
0.8888888888888888,
|
|
"#fdca26"
|
|
],
|
|
[
|
|
1.0,
|
|
"#f0f921"
|
|
]
|
|
],
|
|
"sequentialminus": [
|
|
[
|
|
0.0,
|
|
"#0d0887"
|
|
],
|
|
[
|
|
0.1111111111111111,
|
|
"#46039f"
|
|
],
|
|
[
|
|
0.2222222222222222,
|
|
"#7201a8"
|
|
],
|
|
[
|
|
0.3333333333333333,
|
|
"#9c179e"
|
|
],
|
|
[
|
|
0.4444444444444444,
|
|
"#bd3786"
|
|
],
|
|
[
|
|
0.5555555555555556,
|
|
"#d8576b"
|
|
],
|
|
[
|
|
0.6666666666666666,
|
|
"#ed7953"
|
|
],
|
|
[
|
|
0.7777777777777778,
|
|
"#fb9f3a"
|
|
],
|
|
[
|
|
0.8888888888888888,
|
|
"#fdca26"
|
|
],
|
|
[
|
|
1.0,
|
|
"#f0f921"
|
|
]
|
|
]
|
|
},
|
|
"colorway": [
|
|
"#636efa",
|
|
"#EF553B",
|
|
"#00cc96",
|
|
"#ab63fa",
|
|
"#FFA15A",
|
|
"#19d3f3",
|
|
"#FF6692",
|
|
"#B6E880",
|
|
"#FF97FF",
|
|
"#FECB52"
|
|
],
|
|
"font": {
|
|
"color": "#2a3f5f"
|
|
},
|
|
"geo": {
|
|
"bgcolor": "white",
|
|
"lakecolor": "white",
|
|
"landcolor": "#E5ECF6",
|
|
"showlakes": true,
|
|
"showland": true,
|
|
"subunitcolor": "white"
|
|
},
|
|
"hoverlabel": {
|
|
"align": "left"
|
|
},
|
|
"hovermode": "closest",
|
|
"mapbox": {
|
|
"style": "light"
|
|
},
|
|
"paper_bgcolor": "white",
|
|
"plot_bgcolor": "#E5ECF6",
|
|
"polar": {
|
|
"angularaxis": {
|
|
"gridcolor": "white",
|
|
"linecolor": "white",
|
|
"ticks": ""
|
|
},
|
|
"bgcolor": "#E5ECF6",
|
|
"radialaxis": {
|
|
"gridcolor": "white",
|
|
"linecolor": "white",
|
|
"ticks": ""
|
|
}
|
|
},
|
|
"scene": {
|
|
"xaxis": {
|
|
"backgroundcolor": "#E5ECF6",
|
|
"gridcolor": "white",
|
|
"gridwidth": 2,
|
|
"linecolor": "white",
|
|
"showbackground": true,
|
|
"ticks": "",
|
|
"zerolinecolor": "white"
|
|
},
|
|
"yaxis": {
|
|
"backgroundcolor": "#E5ECF6",
|
|
"gridcolor": "white",
|
|
"gridwidth": 2,
|
|
"linecolor": "white",
|
|
"showbackground": true,
|
|
"ticks": "",
|
|
"zerolinecolor": "white"
|
|
},
|
|
"zaxis": {
|
|
"backgroundcolor": "#E5ECF6",
|
|
"gridcolor": "white",
|
|
"gridwidth": 2,
|
|
"linecolor": "white",
|
|
"showbackground": true,
|
|
"ticks": "",
|
|
"zerolinecolor": "white"
|
|
}
|
|
},
|
|
"shapedefaults": {
|
|
"line": {
|
|
"color": "#2a3f5f"
|
|
}
|
|
},
|
|
"ternary": {
|
|
"aaxis": {
|
|
"gridcolor": "white",
|
|
"linecolor": "white",
|
|
"ticks": ""
|
|
},
|
|
"baxis": {
|
|
"gridcolor": "white",
|
|
"linecolor": "white",
|
|
"ticks": ""
|
|
},
|
|
"bgcolor": "#E5ECF6",
|
|
"caxis": {
|
|
"gridcolor": "white",
|
|
"linecolor": "white",
|
|
"ticks": ""
|
|
}
|
|
},
|
|
"title": {
|
|
"x": 0.05
|
|
},
|
|
"xaxis": {
|
|
"automargin": true,
|
|
"gridcolor": "white",
|
|
"linecolor": "white",
|
|
"ticks": "",
|
|
"title": {
|
|
"standoff": 15
|
|
},
|
|
"zerolinecolor": "white",
|
|
"zerolinewidth": 2
|
|
},
|
|
"yaxis": {
|
|
"automargin": true,
|
|
"gridcolor": "white",
|
|
"linecolor": "white",
|
|
"ticks": "",
|
|
"title": {
|
|
"standoff": 15
|
|
},
|
|
"zerolinecolor": "white",
|
|
"zerolinewidth": 2
|
|
}
|
|
}
|
|
},
|
|
"title": {
|
|
"text": "Naive vs HAC-Adjusted t-Statistics"
|
|
},
|
|
"xaxis": {
|
|
"title": {
|
|
"text": "Naive t-stat"
|
|
}
|
|
},
|
|
"yaxis": {
|
|
"title": {
|
|
"text": "HAC t-stat"
|
|
}
|
|
}
|
|
}
|
|
},
|
|
"image/png": "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"
|
|
},
|
|
"metadata": {},
|
|
"output_type": "display_data"
|
|
}
|
|
],
|
|
"source": [
|
|
"sig_colors = [\n",
|
|
" \"#2ecc71\" if sig else \"#95a5a6\" for sig in eval_summary[\"significant_fdr05\"].to_list()\n",
|
|
"]\n",
|
|
"\n",
|
|
"fig_tstat = go.Figure()\n",
|
|
"fig_tstat.add_trace(\n",
|
|
" go.Scatter(\n",
|
|
" x=eval_summary[\"naive_tstat\"].to_list(),\n",
|
|
" y=eval_summary[\"hac_tstat\"].to_list(),\n",
|
|
" mode=\"markers\",\n",
|
|
" marker=dict(color=sig_colors, size=8),\n",
|
|
" text=eval_summary[\"feature\"].to_list(),\n",
|
|
" hovertemplate=\"%{text}<br>Naive t: %{x:.2f}<br>HAC t: %{y:.2f}<extra></extra>\",\n",
|
|
" )\n",
|
|
")\n",
|
|
"naive_tstat_min = as_float(eval_summary[\"naive_tstat\"].min())\n",
|
|
"naive_tstat_max = as_float(eval_summary[\"naive_tstat\"].max())\n",
|
|
"hac_tstat_min = as_float(eval_summary[\"hac_tstat\"].min())\n",
|
|
"hac_tstat_max = as_float(eval_summary[\"hac_tstat\"].max())\n",
|
|
"assert naive_tstat_min is not None and naive_tstat_max is not None\n",
|
|
"assert hac_tstat_min is not None and hac_tstat_max is not None\n",
|
|
"t_range = [\n",
|
|
" min(naive_tstat_min, hac_tstat_min) - 0.5,\n",
|
|
" max(naive_tstat_max, hac_tstat_max) + 0.5,\n",
|
|
"]\n",
|
|
"fig_tstat.add_trace(\n",
|
|
" go.Scatter(\n",
|
|
" x=t_range,\n",
|
|
" y=t_range,\n",
|
|
" mode=\"lines\",\n",
|
|
" line=dict(dash=\"dash\", color=\"gray\"),\n",
|
|
" showlegend=False,\n",
|
|
" )\n",
|
|
")\n",
|
|
"fig_tstat.update_layout(\n",
|
|
" title=\"Naive vs HAC-Adjusted t-Statistics\",\n",
|
|
" xaxis_title=\"Naive t-stat\",\n",
|
|
" yaxis_title=\"HAC t-stat\",\n",
|
|
" height=500,\n",
|
|
")\n",
|
|
"fig_tstat.show()"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "7789a1d4",
|
|
"metadata": {
|
|
"papermill": {
|
|
"duration": 0.003467,
|
|
"end_time": "2026-03-24T02:23:44.222282+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:23:44.218815+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"source": [
|
|
"### Feature Correlation Heatmap"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 22,
|
|
"id": "3397ee24",
|
|
"metadata": {
|
|
"execution": {
|
|
"iopub.execute_input": "2026-04-09T01:07:52.772138Z",
|
|
"iopub.status.busy": "2026-04-09T01:07:52.771948Z",
|
|
"iopub.status.idle": "2026-04-09T01:07:59.647694Z",
|
|
"shell.execute_reply": "2026-04-09T01:07:59.647259Z"
|
|
},
|
|
"papermill": {
|
|
"duration": 5.713428,
|
|
"end_time": "2026-03-24T02:23:49.939152+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:23:44.225724+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"outputs": [
|
|
{
|
|
"data": {
|
|
"application/vnd.plotly.v1+json": {
|
|
"config": {
|
|
"plotlyServerURL": "https://plot.ly"
|
|
},
|
|
"data": [
|
|
{
|
|
"colorscale": [
|
|
[
|
|
0.0,
|
|
"rgb(5,48,97)"
|
|
],
|
|
[
|
|
0.1,
|
|
"rgb(33,102,172)"
|
|
],
|
|
[
|
|
0.2,
|
|
"rgb(67,147,195)"
|
|
],
|
|
[
|
|
0.3,
|
|
"rgb(146,197,222)"
|
|
],
|
|
[
|
|
0.4,
|
|
"rgb(209,229,240)"
|
|
],
|
|
[
|
|
0.5,
|
|
"rgb(247,247,247)"
|
|
],
|
|
[
|
|
0.6,
|
|
"rgb(253,219,199)"
|
|
],
|
|
[
|
|
0.7,
|
|
"rgb(244,165,130)"
|
|
],
|
|
[
|
|
0.8,
|
|
"rgb(214,96,77)"
|
|
],
|
|
[
|
|
0.9,
|
|
"rgb(178,24,43)"
|
|
],
|
|
[
|
|
1.0,
|
|
"rgb(103,0,31)"
|
|
]
|
|
],
|
|
"type": "heatmap",
|
|
"x": [
|
|
"ret_5d",
|
|
"ret_10d",
|
|
"ret_21d",
|
|
"ret_42d",
|
|
"ret_63d",
|
|
"ret_126d",
|
|
"ret_189d",
|
|
"ret_252d",
|
|
"skip_recent_12_1",
|
|
"skip_recent_6_1",
|
|
"vol_21d",
|
|
"vol_63d",
|
|
"vol_126d",
|
|
"vol_252d",
|
|
"sharpe_5d",
|
|
"sharpe_10d",
|
|
"sharpe_21d",
|
|
"sharpe_42d",
|
|
"sharpe_63d",
|
|
"sharpe_126d",
|
|
"sharpe_189d",
|
|
"sharpe_252d",
|
|
"mom_accel_short",
|
|
"mom_accel_medium",
|
|
"mom_accel_long",
|
|
"vol_ratio_short",
|
|
"vol_ratio_medium",
|
|
"rsi_7",
|
|
"rsi_14",
|
|
"macd_line",
|
|
"adx_14",
|
|
"cci_14",
|
|
"cci_21",
|
|
"stoch_k",
|
|
"aroon_diff",
|
|
"sma_ratio_50",
|
|
"sma_ratio_200",
|
|
"ema_ratio_26",
|
|
"bb_pctb_20",
|
|
"natr_14",
|
|
"chop_14",
|
|
"hurst_100",
|
|
"max_dd_63d",
|
|
"max_dd_126d",
|
|
"vol_ratio_21d",
|
|
"vol_ratio_63d",
|
|
"obv_zscore_63d",
|
|
"pct_positive_63d",
|
|
"dist_52w_high",
|
|
"dist_52w_low",
|
|
"corr_spy_tlt_63d",
|
|
"ret_126d_rank",
|
|
"sharpe_126d_rank",
|
|
"vol_63d_rank",
|
|
"regime",
|
|
"yield_curve_slope",
|
|
"yield_curve_zscore"
|
|
],
|
|
"y": [
|
|
"ret_5d",
|
|
"ret_10d",
|
|
"ret_21d",
|
|
"ret_42d",
|
|
"ret_63d",
|
|
"ret_126d",
|
|
"ret_189d",
|
|
"ret_252d",
|
|
"skip_recent_12_1",
|
|
"skip_recent_6_1",
|
|
"vol_21d",
|
|
"vol_63d",
|
|
"vol_126d",
|
|
"vol_252d",
|
|
"sharpe_5d",
|
|
"sharpe_10d",
|
|
"sharpe_21d",
|
|
"sharpe_42d",
|
|
"sharpe_63d",
|
|
"sharpe_126d",
|
|
"sharpe_189d",
|
|
"sharpe_252d",
|
|
"mom_accel_short",
|
|
"mom_accel_medium",
|
|
"mom_accel_long",
|
|
"vol_ratio_short",
|
|
"vol_ratio_medium",
|
|
"rsi_7",
|
|
"rsi_14",
|
|
"macd_line",
|
|
"adx_14",
|
|
"cci_14",
|
|
"cci_21",
|
|
"stoch_k",
|
|
"aroon_diff",
|
|
"sma_ratio_50",
|
|
"sma_ratio_200",
|
|
"ema_ratio_26",
|
|
"bb_pctb_20",
|
|
"natr_14",
|
|
"chop_14",
|
|
"hurst_100",
|
|
"max_dd_63d",
|
|
"max_dd_126d",
|
|
"vol_ratio_21d",
|
|
"vol_ratio_63d",
|
|
"obv_zscore_63d",
|
|
"pct_positive_63d",
|
|
"dist_52w_high",
|
|
"dist_52w_low",
|
|
"corr_spy_tlt_63d",
|
|
"ret_126d_rank",
|
|
"sharpe_126d_rank",
|
|
"vol_63d_rank",
|
|
"regime",
|
|
"yield_curve_slope",
|
|
"yield_curve_zscore"
|
|
],
|
|
"z": {
|
|
"bdata": "AAAAAAAA8D/dAcMUOxDkP7RAxSADB9s/eBBnsgYF0z+jjF6un9DOPxA7uS2Shsc/VQCOKXHDwj9NXTkEwiLAP4OL0KGKJZU/ico7tEUrmT9Y26JYl2qeP9Qv1YdyFKY/os62Pvfppz/t0JPkfHipP9VFjebzVe0/2S814n144z8B+bMsvTraP1aAI+tSQNI/Y8G9BnxEzT9fvAhzGqXFP/bd4KDJnMA/YaeMxGOHuz+kGmz2hqV3P6FRyntK4qC/EPYhkEVPir8Qv7VYTdeavxh4IlKCvX6/y+GNd6uJ6D9p+qlsnMnjP5jsgKw77NA/U/N7J8y2bT8y5uecTwTpP/5GUWOoeeY/btsSmiSa5z/BuidrsY/NP0C2BCHwSd0/UccThgP0zT9MoX0Cy1nlP6a3t0qJCec/5Avf51HKd7+xqMkYLnuzv0CTJJ00AJM/63S+mjKL2D862zYdRy3TP5bE/2z7PsC/HZPe4o3mvb9/gU8ZFHXWPxuc/rj47MY/etRT6CkRzj8LJ+H7TzXMP7JL3/MNuWq/BR9FO3CXwT/9zPdbr9e9P8mnl2ae7Z8/VcR737u4ez/NhB8qTapiv+1Ikk6Dyp4/3QHDFDsQ5D8AAAAAAADwPy7CNDbcK+Q/hJrVtrbX2j+Gyxyih2DVP+NwH62CBs8/NARCV/dDyT8MMO+wlETFP4Plf6zVT4k/n8SW62JJhz9NP4yvadyKP22wAkqLk6s/VIAEOHZWsD+j3cRBXCyyP4ocaW4MSeM/+AhTZX1w7T9Lm7tR3DfjP/o3Z3A9ptk/tYIp6ll11D//wj1X+unMP36vdOpmnMY/Qj1kvcNDwj+xjW172PqVP/HuNLJrrqC/S9HaMcvRjb97Nn3oX+a3v90CN5SE9pa//e5VGRDI6T9a0V1ltCToPzcrb9IabOA/TfWJqcD8ZT8UYfUv2/HpP3p2rlpS5ek/PljVTvvt6T9QbZl1ZN3bP05ra8yELuQ/KDvy40lb1D8sL2fnyxLqPxPlO4yFLuo/Y2aACERfn7+TyGjEKXq+v86wDk1lt5g/k2mHbJk23j+J6j2OP9PXPyxrk/Yuvb+/chYygVZywb+upCB406zdP6NchnWONs8/8lAOIePK0j/X8D8flNLSP7tg4U/4BZC/N++y9gU0yD9KYj7G6/TEPwKL7tWlNaM/C6ty5NiWdj85Yf2S7EYvv6mL+I9m7KI/tEDFIAMH2z8uwjQ23CvkPwAAAAAAAPA/gh1Ne0+05D+I0F5KGsXgPwJXD1bLbNc/YItKy0+E0z/nWJox7I/PPxazgcwYvYk/prgVojcujz90VXOVA7O2v+qrxFSK75Y/ORe1mHDwqj+CBNfQJvyzP2PQ4AACY9o/KfI3i32m4z9CiObuhX7tPyR5x6fSjOM/50j3Dh623z/8pJ4UTO7VP2/eabG4pNE/5D/b9mOsyz+VWjm1/Lp6PxO8LXdhpKC/2dDlbWptkb/hONofoHDTv+4/mSjRorm/uqOBGtT05T8PW5j4st3pPy3DEWmt3uk/xnH04VmZTL8XMsqrI43iP69bZOIQe+c/fA9D4Oww4z+JY8GuR1zoP3gqqWXuXes/ymXFC8Fx3j/I9c/stJTrP+imfdbzzuY/TP1EgAirub+0fm67/wq0v1+zczxWlaA/RgLMss3v4j+UIvEj2x3fP7Au+PAr1rC/mJ5qU0Fnwr9ojG/24qTiPxJQKI2hhNc/bWkcQKMB2T+x8kP0XHXaP1W+aT4a74i/AVdgd0Bb0T/G/iXqqkjOP85gTvgtV5o/UmWtM58skj8HCqZp8qp7P33EkHA8jLM/eBBnsgYF0z+EmtW2ttfaP4IdTXtPtOQ/AAAAAAAA8D+3MNb0cKvoP0TLuwVhpuA/fFXM+Pa+2z984DyYbrjWP2PYCNenN8k/cUyNUVi40j+x1eFrtr/Cv61+pUcjxJ+/S98Erd+Fmz89dCyqPDKzPzd9/sIeM9M/YzGgij4Z2z8xEVcAnQLkP88j/044Ye0/X5df/Xnf5j/yn1bGJv/ePzpXb6HLR9k/fWSaB9ke1D9Ah4Rx+kjdv/Vg9aIPZJq/3UUumMpRob+hTr8t+bXUv9+JzMQquMq/n9VoOVm+3z9pcVeHKfjlPyUDKX4IEuk/Gq6TAB12oD+6b/3UGGHZP7otXRXZu98/y44myZ+Q2j8ZMBIECRbjP/8Iz8AQguw//x3CzFwQ5T/VaoONElzmP2NC4jT2/t4/9L8C79i5wr8sJ4PgW3t9vzsgjG3oQ34/LC7V7Kct5T+QhzTK05DiPwfX3lRNloW/4gB6IJoFuL+QiUPXn7njP7nCLNEJ2+A/f6YlQH6U3j8Hp0gAsvHgPxRssY9ypHE/ex5eMIaS2D92yv4jAsPVP8dMUU/UhpA/rGgFvQ08qT/1gMmaXjySP5dLyzSowLo/o4xerp/Qzj+Gyxyih2DVP4jQXkoaxeA/tzDW9HCr6D8AAAAAAADwP9qYKPfw3uQ/woAbXaMY4T+bvBG1mCTcPzHSo0FM/NM/IL7/ZcyQ3j8VO2PVk47Cv47i60WypLW/3KmIS9REfT9R+YZ/pW6yP+T+N7L6NNA/P6cMmPV61j+vmRv24obgP8rerIjuVuc/C1j0ToYU7T/bupWUACvjP2pDVgkV5t4/gj8fdVXh2D+zLbOyNxrpv16d20M6g5q/K87t3o5Fo7+MQiYeBuDIv+gmGyDmrtW/oZ8npgyp2j+z8Wdki77iP+2BRoEeUuU/pX6K54F4oj8ZC8bIh2LVP/31ZQEhRdo/cK45wtph1j/8R7m6z+3eP8pdMwvRBec/o0tPtu3I6D/FkRtO5qPiP97rjZlOw9k/avHkmJ9vw78Gsk1Yon1uPy3EUFpY/pC/17wtex0/5D/RoMqeqsnjP7/tkS7PyoW/3NSnTr0cqL8uRhximt3hP1vywqTWzOU/GHn2ZOq24D9CjKcJ3VzjP2LBoketsqE/XKRnDd+E3j9OaeOEmcDaP2knCq4J+Y0/b0T2I7qRsz/a+WomQ/KfPyMlihucgsE/EDu5LZKGxz/jcB+tggbPPwJXD1bLbNc/RMu7BWGm4D/amCj38N7kPwAAAAAAAPA/L2sz4u3/6D9gAGeTfhnlPwM+H1D6gOI/4MEAt1Ds7D+SXgNTvz3Fv2CZltbH2cS/SY3pVw2tvL91Uo7ureN8P4FtU4Rplck/tA5iBNLx0D9BGPgovrPXP1ANmfADOOA/jVv5WBL+4z/LWhkV2M7sP4ERQ54ihOY/Psdwae3y4j/6Mb+vmYDgv88H2sSfa+a/5fSeDirMrb+/SUR1XHGqvznWHzDCxcq/DByxsebY0z/Po7QefUHbP3x6CHPRKt4/MEKK8dFSgT9UdCnwngbQPwGl0binbdM/UL9hFoJR0T/NgteXuQTWP8HccVEaZN8/Ql7ITydm7T/k95lnLHLaP1wLxFR68tI/dmBpBVF2xr/ki1bkDA+BP8nGIhNZVa6/mgB/7O4N3z/qNZvqYCTkP6IaVM4ZaoS/A1v4uioAkL9/84zojnXZPxJ94+obw+E/scgY2o5k4z8nB3ANLUTmPyRlheODX70/SVMGyU4G5z/X4B5H0b3jP7mkSXjiv3U/IJ5S4zxguT9aw7DYo+auPxt+SnMKt8k/VQCOKXHDwj80BEJX90PJP2CLSstPhNM/fFXM+Pa+2z/CgBtdoxjhPy9rM+Lt/+g/AAAAAAAA8D+hIqNl2b3qP/oWJDSyseg/tGeYM5+f5j+4qZSEIlDGv16oiDwuesa/YpCmXL7jxL8j8UVpRv6tv1MoLqSfasU/2eXIGjbGzD8pDO6CuV7UP/piHhDEqNs/02M4wka44D9KmSRrk2fnP9Bd4FW8rOw/p4PggeL15z+cTHOx8g3bv4pOpD1KeuG/vHE+qZVn27+Df/DnG9ekv9RulsN17ba/feNKhgYF0T9rwbQDgDXXP3CuWC0f3tk/VJN6xJubZr/0G+5vzeLKPxU84oNxYdA/feHjNAK+zj9HAsgH4QTTP7Nx7UiTHdo/e2ABUMXA6z9z1AFbIurVP76SA3+7E9A/lEewUSEtx7+Qaf/gc6ltP/8POOWf35y/xmwhLTdI2z9crlLj4HPhP29IBuxpQIe/0W+ffXVWjb8B1ymEegjVPzrvQVCgRuA/nmtjbgds5D8/iyeDSUPmPwlk0JV6jrg/OXtfAf574j+f+++ZZWXgP+cst9E3GH0/Mpkmi+lUtz+832hxbQK0P0lA46THYcY/TV05BMIiwD8MMO+wlETFP+dYmjHsj88/fOA8mG641j+bvBG1mCTcP2AAZ5N+GeU/oSKjZdm96j8AAAAAAADwP+3gfQ8Ol+4/hEOfUztJ4z86RngXB2nFv2tAVAe9Vca/CYlARVngxb/SSMe8am+9v2DRCLY268I/j4qd85AZyT8g0TV4MTDRP9K21X/2V9c/Cg4EUz8+3D9xB95+kxHkP+vrQUBDmOg/YboROghx7D+uO3a4ic7Wv66iY7izXN6/t470gwkI579mPSMj87CTvzn8PtmHZKm/NuKGJKT3zD/ah6bosofTP+bmar0KsdU/v5DHe3y5b79mRzrYKzTHP/+SlSNAxcs/IJuV9Pnxyj+AL3bBLtTPPw7wLPZUK9U/D0NjnSRa5z+bTKI8zerRP+FmbZZEQMs/K7875J/Wxb9yvoPzEHhFP5KiEMvV2oy/j207yhUi1z+AIVg6mbzdP4Pe9fOwaDg/hjrfrE3fcz+JCB+LDiPRP60UPtFiY90/KheUu02l4z+6SKetEibkP5XPdXOC9bQ/g1YXmc+Z4D/3FdnoCNndPwxKYwTRZnI/eOY+Pj6DuT9NEGH0ER25P3k3gQj2tsI/g4vQoYollT+D5X+s1U+JPxazgcwYvYk/Y9gI16c3yT8x0qNBTPzTPwM+H1D6gOI/+hYkNLKx6D/t4H0PDpfuPwAAAAAAAPA/WAOnhb8H5D+QJLRhhr3Avzhs+xqsscW/QqQZ/b1hxr9wT0je+gjAv5/KUry0Cac/JPYWlCZopj8jcudcB76mP1+gwC1ogcs/mmX9U9mD1D/8yuhRQ6LhP0Qhjpt60eY/tYxASmRW6z/3q3kMJ7jXv3r3GbImPN+/bt/785Tj578SQthVkXKxP+hJ6ZiN1JS/FJH2P7UFrT+kPyMtR5y7P2aKuzZ0k8I/76cch+x1hz8Je5pq8uujP6023qEPiKU/4rXpKv49sD+BrQCqhtywP0wEe1BVQ8A/awJQ4s/E4z8Up4ptn/WyP+tGf4cfTaU/fsa1FAGXwL+hFvC0+4SJP9Em0GcuxpK/ooqPt7QFyz/YJkJ8wQTWP6klNYfHdpY/jvfhAEBfqj9TwV4rsDe/P6t22VjbL9g/DIsVLVGE4D8GABFDWHHhP5k5gcQHiLY/NIuuUYey3T+Gz6X+nePaPyBgOUizdoQ/kKM1nz/iuD+YJ50h7Yq5P/HL6eFwHsE/ico7tEUrmT+fxJbrYkmHP6a4FaI3Lo8/cUyNUVi40j8gvv9lzJDeP+DBALdQ7Ow/tGeYM5+f5j+EQ59TO0njP1gDp4W/B+Q/AAAAAAAA8D/okzGgVwy+v5c45+MDpsS/FAo7D+u4vr8/qC7XhXtkv5tWVVJ6Gqc/o9tLpTCBoj/LuAZfDOafP+1qOD3WkNI/65v2sJaK3T/E1WBgeUHqPz3kPLRwYeQ/At9g2ZVJ4T8Dx+oUtyjiv7AlgKHFVOi/0AB7OhHbq7/dZCpTOlqzPwqEmQ07fMa/p1N5qkDxqz92AeJtU7XAPwyvxVVAscc/hvLuX1y4kz9xVrEezKmiP88yjjJ5kaQ/NuyHdhp5qT846DjeQYSxPzolcFpLOsg/oX8aYVyr6D/boFuTNGK7P5rpykrvlaM/rmTCZ80Zv7+M2ViRBLugPzfGf/rHz7K/xY4QhScF0T8rllvJohrdPxDSp51n7JI/J0w87njTqD9JsQU4LsPHP80hlPGf6dw/EkzBUNhs3j/vcpnFZPziP9fGdwiJ4b0/HodOsgTt5D8tPsh/8S7iPxmOeKLB6og/SGMbM48ZuT/N0CU/pS2wPyI0Bpzgxcg/WNuiWJdqnj9NP4yvadyKP3RVc5UDs7a/sdXha7a/wr8VO2PVk47Cv5JeA1O/PcW/uKmUhCJQxr86RngXB2nFv5AktGGGvcC/6JMxoFcMvr8AAAAAAADwP+Oafb6vfO0/qjnHeArI6z8MmUAWThvqP0fZQMfExa+/dLcIFt75vL9BErmrQODNvwkP3uS1dtO/cceximCK1L8y/d5k+b3Wv0t8l0nS8de/FE6Zh/VH2L9CPyQXqna1P3CJ6vHmHaU/S/nF7CekmD8l3RIb733SPwlZuzursM4/8MqzyA4lwr+WQ2rMbNDNv1WyjK9tzca/WpAcdAI+ob9HOHtEuYS4v9VsNzagocG/zhGTSO7kt7/W3oSicjbOv/s+6qT3O7+/9a66yVmyyL9KVPeZrY+rv2MjVmT5JsG/81h2boTH7j84OtIRKyGVv1vdM8jsqaE/QPwNm01j4792EsFVZVzkvy+PemzjKJC/eg5d5BAgqz/0owrWgGPMv+jHOswzJce/FXyCmJP55L8WsdMBF53QPzCaBMoN1a6/F/Q+cx6xqr77UOoS/CDIvxf08QpySOg/DXXjBIgnZD9LJ1n2X2ikPwkmVogfJ6U/1C/Vh3IUpj9tsAJKi5OrP+qrxFSK75Y/rX6lRyPEn7+O4utFsqS1v2CZltbH2cS/XqiIPC56xr9rQFQHvVXGvzhs+xqsscW/lzjn4wOmxL/jmn2+r3ztPwAAAAAAAPA/pJC9UV1m7j8z2GcjxF3sP4zyu5Ui6ae/Ov3NdWNIsb+6VbppaUq9v3ZxeAfqQ8m/JOZb1Ez30b/inU4mElnXv1E8b18cvNi/83vhXGaH2b/M5eoY54e3PyxyTT7dz7U/hAfe3xZDnz82vRCRsnaovyNfMhIeR84/iB/2h/zGtb93oXHC7wzCv11qCkkDUK+/BTZa9uMQtr/lRMgFYiWtvyS+ig0XF7O/tWAQqNlVrr+jBsnKGhW8v2g760K3G06/WjLmYVzjxL8ob9alYb+gP2A5fiqKSbO/WTrnywM07j+pm3+eu+1qPy9BwEPcfKM/f7Y7J+gP4b8O92qFhzTjv0NMrl4WXWy/kWwFQSbRkr+Y5m459nrDv7WCgBIRzsO/7jhzQpp95L9y9FBfUbzTP72iftSANrW/ARC885nJcr8TcJVueb/Jv/7kFQj5n+k/a2PyJcBTbz/AEpoP8BmmP5zcAvsOka0/os62Pvfppz9UgAQ4dlawPzkXtZhw8Ko/S98Erd+Fmz/cqYhL1ER9P0mN6VcNrby/YpCmXL7jxL8JiUBFWeDFv0KkGf29Yca/FAo7D+u4vr+qOcd4CsjrP6SQvVFdZu4/AAAAAAAA8D9wmNChLzDuP8ZXdQUT6aO/cU09k2N8qr9fRlKUoZe0v4eZb9/gM8G/ejINGgyWx78QfF7FXSPVv/umU747ydi/attsrTYL2r/vic1cdbxkP6IlhHpyU7k/fdMfiF3+sT88qwAm0jyzv5hv+aQIO5q/af6RlySBsL+pnf7nu+q5v5b7Qmg/KIa/Z6Usrq4Ju78y6Dw0coClv6V2g7r4iKq/hoXwhmQ4qL8quU0e/0m0vzSXNiPCoac/j86qzbL3ur8IexvYn2OwP+lsdChgc6u/f/OhOLfM7D8u3tHSvSWVP5TdgEhZqnG/FzxxP2ID3r8+xk4q5EPhv5d/nHn972M/LH8rnmSQk79NsAzq+6+8v0yfBtXaL7u/QLO+ju9K478MGAPPeIDXP2+P3VHaBbK/9lUxvKTPPz8N/QugXBLKv6rAwhItI+k/HIf6pJ0FmT/eddIyfcypP8/K0hcpfLU/7dCT5Hx4qT+j3cRBXCyyP4IE19Am/LM/PXQsqjwysz9R+YZ/pW6yP3VSju6t43w/I/FFaUb+rb/SSMe8am+9v3BPSN76CMC/P6gu14V7ZL8MmUAWThvqPzPYZyPEXew/cJjQoS8w7j8AAAAAAADwP2td20z1Q6C/fLnlUBCupL/nevaHvsKrv50gvBL9+7W/2QAAiwsMvr+nSXtKLtXKv0qzaiuq2NK/begO0tTV179/0BLJ1uSrvz9DcZGNWWq/6rwRBXvOuz8DxMKxa9ixv1OcCb2eE7S/ZRac0yPXp78FA07TpHixv0V4mg4T0Jc/HHNGMdNlvL8flWdu5Yafv/riQMnheKG/cWUUObusob8MG7SRR5qrv+MNNjLVA7Y/vm3YdiwpdD/Pedu3LWa3PwinF4A9CaO/i0Lc8pk36z9QZMn21h6cP98lLiBAV6K/zkXswmYP2r+TfJhB2CHdvxZCHv6er34/cdEyAN1qgr8sy765nnWzv9UKGaKGBbC/aBnPennX4L/iwH8s03jePwHQ38R+M7C/whJVibyUlz+9zKHfwPbFv8lmg5TNk+g/ljwsiUD0tD9PHSo16gS3P7efd2qrUsE/1UWN5vNV7T+KHGluDEnjP2PQ4AACY9o/N33+wh4z0z/k/jey+jTQP4FtU4Rplck/UygupJ9qxT9g0Qi2NuvCP5/KUry0Cac/m1ZVUnoapz9H2UDHxMWvv4zyu5Ui6ae/xld1BRPpo79rXdtM9UOgvwAAAAAAAPA/ie8f/XZd5T/8rZRbgGPdP1CSrlTSqtU/amO+8lV20j8tcBmjZkvNPy6KI+cTqcg/lZuVqg7yxT/57DPX1yyNv4oabdyn0KS/sGUEu6fBmL8atYG2seOkvzfv0GCclZy/NJ7LN2m66j/LSN0tuNblP1jdSava89E/Hzb0j4Zenz+NKfzjeIDqP9Jl0n5b/ec/KG+STD1S6T89WEjOX5nRP3fic+1n4tw/lZXAJxEl0D+L0x7hDY7kPyWVMHkMoug/V9vAUBntuL9Lf6F9jfm7v5+65MknrXc/7jFoBt9V3T8S9nY1K1nYPw81Bl9+6r+/U+EUHYg9vr/Lu1PJarPaP25vZfl7zs4/PRpirRWC1D8cjORu95vHPzTvEpDc6oK/c4HBFvK+wT/qJCWbZgzFP+NIUkHbCai//doH7PKedT8tcO6TFhtmv1uRwN/3G5s/2S814n144z/4CFNlfXDtPynyN4t9puM/YzGgij4Z2z8/pwyY9XrWP7QOYgTS8dA/2eXIGjbGzD+Pip3zkBnJPyT2FpQmaKY/o9tLpTCBoj90twgW3vm8vzr9zXVjSLG/cU09k2N8qr98ueVQEK6kv4nvH/12XeU/AAAAAAAA8D/+5LK3UYDlP2Zg/r1TQt4/PsN4ZlKS2T84z1WL3XbTP5PpwNKAgdA/rS1MpCMOzT9/yuYhNHhkv7bvEO+UtKS/zk0XVW/Hn79iIvoIyiu/vxadaPr2J6y/5zQjp/JW7D87npobM9bqPwHFq4oD+uA//Bo8PFikpz94C/1w2XfrP63a9Ls/qes/0zD4lpHN6z9j83AnEe3fP1wn4euJ2uM/b4pLgLH/1T/OVCi7SfroP4dWsvW7/us/bUm8zYmQw7/kToxx35jGv7jWIBfVQoo/Txwlatcv4j95dnJbCnjePzbfjWGn4L6/bMQJdV+4wb+nUZ/H5Q7hP0H58C/UJdQ/n0NFepDS2T+OYTrBg2XPP3RGMKVcT46/rAMrwbhGyD/GydiLW6nMPzhU2zbI6K+/UGcLj1p7Q78PLv5O4k99v9iegqEaypw/AfmzLL062j9Lm7tR3DfjP0KI5u6Ffu0/MRFXAJ0C5D+vmRv24obgP0EY+Ci+s9c/KQzugrle1D8g0TV4MTDRPyNy51wHvqY/y7gGXwzmnz9BErmrQODNv7pVumlpSr2/X0ZSlKGXtL/nevaHvsKrv/ytlFuAY90//uSyt1GA5T8AAAAAAADwP9XC2sMACuY/nz8TFBx94j+8SX2jyfraPx7w5NbcAtc/KWdlKD2h0z8L4NXkkVF0v1scE6nGQaK/G3BgCAYYor95KX+IhUXVvxetO/5vgr+/hZYrb9Aa6D/1XiQJoW7sP/a28AwuWuk/XyHkT1SKtD9hb9wm0jDkPzS4gcYJ9+g/l1frVJrf5D8zeTb9C37qPwzCFBA53uk/kI2QDILK3j/NZlH07NnpP0xnZ928Z+g/+0gF7TUazr9zQJ4x9vO4v2Ropd9lUJY/iaY8rWfb5T95Nlbpo9fiP2EhCUAHia6/z7GIk2Nxwb/Puypz25nkP+SsLsLSuNs/qOc+bgAk4D8+tj04xqrVP4OEL9039pW/1GRfKkPU0D+5XIEqWcbTPzfXHcF/7be/p2Mi0PdHhT/v7WmejWhYvybczYR0aa8/VoAj61JA0j/6N2dwPabZPyR5x6fSjOM/zyP/Tjhh7T/K3qyI7lbnP1ANmfADOOA/+mIeEMSo2z/SttV/9lfXP1+gwC1ogcs/7Wo4PdaQ0j8JD97ktXbTv3ZxeAfqQ8m/h5lv3+Azwb+dILwS/fu1v1CSrlTSqtU/ZmD+vVNC3j/VwtrDAArmPwAAAAAAAPA/LhMxzceu6T8T7YoHSUriP9x0Dv8PKt8/ohFBujpz2j9jnZevB1fcvxSZWDGTi5u/YNuGdN2Cp79i22kn4U3Uv9GJsGM9ec2/5KxFE+ij4T8z99THjjjoP17dim+HX+g/R6DioB/MvT+5hRCcVV7cP7nNircci+E/zFW3keZw3T9+dTwP+QPlPz3ayBBsbuo/4Vvr0RR25D+Bsm84rfvkPxBt+ZzzI+E/bLjukyFi0795HzMZOn2UvwXna79KQHa/PAV7bezS5z/Ka+Xto+7lP9o9eNH2m4a/ZGmciE0str8M4cEQBFrlP7s73X9aveI/qlECc0ck4z8mZcm9eHnbP5bX8xgPXI+/niEe3i4M1z8Aj6UXOzfbP+2Ss6F/AMG/wWWXAD+yoj+yyGW46QVnP5OKr44IpbQ/Y8G9BnxEzT+1ginqWXXUP+dI9w4ett8/X5df/Xnf5j8LWPROhhTtP41b+VgS/uM/02M4wka44D8KDgRTPz7cP5pl/VPZg9Q/65v2sJaK3T9xx7GKYIrUvyTmW9RM99G/ejINGgyWx7/ZAACLCwy+v2pjvvJVdtI/PsN4ZlKS2T+fPxMUHH3iPy4TMc3Hruk/AAAAAAAA8D9X2r9MkmXmP3HLaF8ivuI/9gyLG3Tb3z/5lRldN13nvz4YdeCOEaC/BOsoqwldp79J1ByTop/Fv/KpXBfh3ta/RSdgrAr83T9rCLuGkOXkP0N3+wih6OQ/E55Fc3qouz9wy4Jn6BnYPxHkbalidd0/FOn6kEgH2T/S9fD/GVPhP5aujetzceU/sHAqmYWQ5z9cJ2uPKXPhP9ljj62w5Nw/UpDJ7XAk1b/xhtaLpSF/vwXTja6b45+/Z8dwjCH+5j+7q9a2qknnP4Y02gQGVI+/jigbGQBzpr+nuTlZp2PjP5Gkh+BBcOc/Whlk3Kjm5D/7XAJW/MreP0iD8dHmLIg/gSsVGEEW3D/sfTyiSo7gPz3orb9BxcS/vQhjRGAKsD9AfaMaEteIP3BXDahpILo/X7wIcxqlxT//wj1X+unMP/yknhRM7tU/8p9Wxib/3j/bupWUACvjP8taGRXYzuw/Spkka5Nn5z9xB95+kxHkP/zK6FFDouE/xNVgYHlB6j8y/d5k+b3Wv+KdTiYSWde/EHxexV0j1b+nSXtKLtXKvy1wGaNmS80/OM9Vi9120z+8SX2jyfraPxPtigdJSuI/V9q/TJJl5j8AAAAAAADwP4d4T/fqEOo/V1xBRayQ5j+RVNhadSLev+8NHcaOyOS/r4Ro0K7srr/iO08tFqalv5GjyHZcnse/t/Y5ZtWd1j8hkAJf7OPeP7+fhgdgkN0/wsF9AVAZrT+uM3EuGkbSP1U504efA9Y/ibeLuglo0z9AsCKNzzTZP+N2s/l3Ot0/bVg9EhXU6j++dN8nK3bYPzU8o2gihNU/a1Ot4m6c1784n+gbF7Vuv/c6FURWjrC/e8YtYurW4j/6wD/5J0nnPxTxXE1FP4m/x/3P2NKmjr/5tK+c8r3cP3xqP3JP++I/KlPRdpth5z8WkU4lc1vhP6fRemiEybM/5xe/w2QQ5D+SRAzuSS3nP0jsh2D4rMi/PcaMx7sdtj/Ag8NPyUqkP6esWRku5sM/9t3goMmcwD9+r3TqZpzGP2/eabG4pNE/OldvoctH2T9qQ1YJFebeP4ERQ54ihOY/0F3gVbys7D/r60FAQ5joP0Qhjpt60eY/PeQ8tHBh5D9LfJdJ0vHXv1E8b18cvNi/+6ZTvjvJ2L9Ks2orqtjSvy6KI+cTqcg/k+nA0oCB0D8e8OTW3ALXP9x0Dv8PKt8/cctoXyK+4j+HeE/36hDqPwAAAAAAAPA/HlEsmteU6z9k5DK97lnYv3nqp/arTt+/XTU6wQsI2r9/3KDetkehv8Hofc1OxLG/Y0ToRKRJ0z9WMfSKUynaP3qSc1bFztg/UlFQmNGLpD/nI9vSV4rOP5Fqvdbtb9I/ISjMoLv90D/5f7CWwJDVP0nPBuvrotc/8alOUvTi6D/kYya6J5TTP5WPv8AZHNI/y6xBgTim2L9lmy7v3552vzALDNQNiqG/fUdMME8F4T+OFFx3OOPkP68xJxlT3Ye/BkCO11Lmh7+JSM0Tv+jXP5v07uzDOeE/g9oeOQEE6D8s5kU7DszgP/bxfyBeRLA/s19llJXk3z9Sl1LBDzDjP+/ppezyH8q/DF8krAIKtD8d4drKIdysP1nKzVhM3ME/YaeMxGOHuz9CPWS9w0PCP+Q/2/ZjrMs/fWSaB9ke1D+CPx91VeHYPz7HcGnt8uI/p4PggeL15z9huhE6CHHsP7WMQEpkVus/At9g2ZVJ4T8UTpmH9UfYv/N74Vxmh9m/attsrTYL2r9t6A7S1NXXv5WblaoO8sU/rS1MpCMOzT8pZ2UoPaHTP6IRQbo6c9o/9gyLG3Tb3z9XXEFFrJDmPx5RLJrXlOs/AAAAAAAA8D8xkNx4JeTTv21jm2sfCtu/Eat3zRQp5b9vpxnURF+Kv3gbWpE7S6K/JkrbmF2d0D/1bf6LrUXWP2ZZSIqhdNQ/tGNhSG3fpz973vAhRozKP2vAHlA+bs8/XgHFoUfyzT+MOcJxvDfSP5mFSxdwnNI/OmsnTIDR5D/fZmt0fPDOP2Leg6V7+84/zrUXJ8bK2L99AtY166iCv1cJvFztd5K/NBVLS9kz3j98loSrSJDiP5OOkHaFLUa/sIX4zr9Kej+8gg3cKt7TP0zdrrS06t4/FiRlOOL85j8G4BiHrmXcP9SbPfepTLA/KNyfCcZh3D/QIvSKD2vhPytoRijiAsy/2H38stogtD+DspwaTYuyP7WTEhTDcL0/pBps9oaldz+xjW172PqVP5VaObX8uno/QIeEcfpI3b+zLbOyNxrpv/oxv6+ZgOC/nExzsfIN27+uO3a4ic7Wv/ereQwnuNe/A8fqFLco4r9CPyQXqna1P8zl6hjnh7c/74nNXHW8ZD9/0BLJ1uSrv/nsM9fXLI2/f8rmITR4ZL8L4NXkkVF0v2Odl68HV9y/+ZUZXTdd57+RVNhadSLev2TkMr3uWdi/MZDceCXk078AAAAAAADwP8WyldHMNoA/aE8p2yc4pD96MQS2MQyTP2IYsT1f89Q/Pi1BTMERnr/jWftPbAvDvwaDTOJ9ANC/GXnpMkSvqL+VNOj0Akp+vxHKsjDeo4a/Vv9eDye8hL8Rsvf+OAazv2vvP6Ff19G/qCFwOhZv4r/91BgVWCzAvy7rVy8o9Ye/uFMXi5NPtT9kJGe5v9Krv7LM4HyPQaQ/TQdSFRcH1L9ZxRFXDlvYvzDklJtKbZ+/neV+4lC5pb/KzY/cqFDQv2+XjiF8G+K/OHyu/VSP1b91rq5bOufcvwwkuYlm06W/2LiFUzDd2L8g4GVInB7Wv7vMbq8fHIW/DKcetbDdsL/p9+lGeD2Yv4gLEin5+7y/oVHKe0rioL/x7jSya66gvxO8LXdhpKC/9WD1og9kmr9endtDOoOav88H2sSfa+a/ik6kPUp64b+uomO4s1zev3r3GbImPN+/sCWAocVU6L9wierx5h2lPyxyTT7dz7U/oiWEenJTuT8/Q3GRjVlqv4oabdyn0KS/tu8Q75S0pL9bHBOpxkGivxSZWDGTi5u/Phh14I4RoL/vDR3Gjsjkv3nqp/arTt+/bWObax8K27/FspXRzDaAPwAAAAAAAPA/XBSpgiugoz/p2ax2Rv25v8m+QvtCj5i/3KFZryk/pr+JNHmAJ/+nv+X25FFt+qa/jMJ7HR1hlj9IkFEfyQSjv3Wa2wWWfKS/PiJih14Rq7+VMHxiamqgv7KktvqQhJ6/QHp/Kwoa4L+qosyaRc+mv8iDKgRWD6O/vkQAtJH9pj/KY/T4/BVVv9sDOn1sL7E/eqvSkSVvsb/a0u7ZTdPOv+UZ0nPCf1g/O75cHlrqmL/rasfOCw6jv9yKJ3kxBsK/ArUMPuV00798shL9A6zav2O86BmoHrW/X6p7dw8c4L/+HMQMyI7cv6enIZc6A6O/S8fHhbfOs7+dlkb0LrOxvyXgeC3hWcS/EPYhkEVPir9L0doxy9GNv9nQ5W1qbZG/3UUumMpRob8rzu3ejkWjv+X0ng4qzK2/vHE+qZVn27+3jvSDCQjnv27f+/OU4+e/0AB7OhHbq79L+cXsJ6SYP4QH3t8WQ58/fdMfiF3+sT/qvBEFe867P7BlBLunwZi/zk0XVW/Hn78bcGAIBhiiv2DbhnTdgqe/BOsoqwldp7+vhGjQruyuv101OsELCNq/Eat3zRQp5b9oTynbJzikP1wUqYIroKM/AAAAAAAA8D9gMlDVdkpmv9gUKFhxaL2/HdrRNfK/oL/8N5nE3Lyiv7OfbB6V6Ki/LLk7+BU3kz+ZJBFt2C2Xv35sdqfKrJi/1WOJra8qqL8TqZHYu9mdv3lHSt6PZZu/uY+HZ0Rzyb/Ns3/2sWWYvyvPE6pNspq/CUcAFK2RkT/TsQ6pE8eCP4ULSuK79qe/ySXmU2jLo7/gm/vg7butv6Jo7iBFr4u/jhtvD0ADlb9B9PfA60uTvycjm2hj9cK/SKJbUs8gzr+jgyVlVaLRvz4ujDuf+5C/380Dg985uL8ZCHkejPS0vw1i8v57wqe/jbXp4aQLq78i+IRGu724v2wMZT+alae/EL+1WE3Xmr97Nn3oX+a3v+E42h+gcNO/oU6/Lfm11L+MQiYeBuDIv79JRHVccaq/g3/w5xvXpL9mPSMj87CTvxJC2FWRcrE/3WQqUzpasz8l3RIb733SPza9EJGydqi/PKsAJtI8s78DxMKxa9ixvxq1gbax46S/YiL6CMorv795KX+IhUXVv2LbaSfhTdS/SdQck6Kfxb/iO08tFqalv3/coN62R6G/b6cZ1ERfir96MQS2MQyTP+nZrHZG/bm/YDJQ1XZKZr8AAAAAAADwP5gYbqdSVLg/Az5yzbDzw78ZmARBFbDQv/WJT/JY7dS/k5JMZpl2xz8Xa4kNl9m5vzPu8lhpBsa/zf0ZJNWmt7+iKf8qynbWv8WEmYT/K9W/HzzzG+qpvb8dEI0Ip4LNvxMGBgmsl8S/6gOwGNJvwz/ZHdKvTpiyv2/M7v/QyXM/SpUECSTKz78kRugBWbPEv4ZvZvBLuqC/FNMXmK57yT890QC/aJHMv3OgJpdDFru/RoT2gMWsu7+EgnbRaljBvznrusMAGKI/cqP+8+CiXL+7b0fsPtpPP74MynZlsIK/STQao2bKeb/mxQmdOriQvwLsKKD7iqG/GHgiUoK9fr/dAjeUhPaWv+4/mSjRorm/34nMxCq4yr/oJhsg5q7VvznWHzDCxcq/1G6Ww3Xttr85/D7Zh2Spv+hJ6ZiN1JS/CoSZDTt8xr8JWbs7q7DOPyNfMhIeR84/mG/5pAg7mr9TnAm9nhO0vzfv0GCclZy/Fp1o+vYnrL8XrTv+b4K/v9GJsGM9ec2/8qlcF+He1r+Ro8h2XJ7Hv8Hofc1OxLG/eBtakTtLor9iGLE9X/PUP8m+QvtCj5i/2BQoWHFovb+YGG6nUlS4PwAAAAAAAPA/iaujXIr2s788lmptZP7Cv2CbtvxVm8W/ovKi4pbKtT/hmah+cOqrv+2goeu827S/3PiglXJzp79dgP3/tYq7vzfv6LPLbcW//faPdJPYzL/VddTRP/27vw7J85T5ZbS/xVDTL/0cyz8B75yubxWsv504g9mBGL8/5ZzALNoO0b9pYdZnbGPQv2APNhHlJJy/LY1pKZYkbr8J8tAe+pfEv+rDRnVHE8m/w56gsrvuxb+TrOkF7VPHv/HdchKiwLG/MD+idiVFlr+OvOeFBkCXv/5+XXU4arg/vKK+hWnvsL+koEZSuJ+lvzYrRvpTHa6/y+GNd6uJ6D/97lUZEMjpP7qjgRrU9OU/n9VoOVm+3z+hnyemDKnaPwwcsbHm2NM/feNKhgYF0T824oYkpPfMPxSR9j+1Ba0/p1N5qkDxqz/wyrPIDiXCv4gf9of8xrW/af6RlySBsL9lFpzTI9envzSeyzdpuuo/5zQjp/JW7D+Flitv0BroP+SsRRPoo+E/RSdgrAr83T+39jlm1Z3WP2NE6ESkSdM/JkrbmF2d0D8+LUFMwRGev9yhWa8pP6a/HdrRNfK/oL8DPnLNsPPDv4mro1yK9rO/AAAAAAAA8D8LbUK55t7tP4uASFog3OE/Ywzb4gUprT/4CVsdgP7tP3IHbqJQ/O0/LCLtfOYz7j8o/BjJYVThP6MJFyURZ+Y/N9aentG82T8IP8obh83rP1HQUQ9asO4/hHA652qaxr9b8ZPyY7rEv3Nu8FFpFZU/Mqe//S0f5T9BDeTSBcbhP4Oclx3kMr+/Cc5LbVm3wr/wcx5m+ALjPyieAugxZdc/wTReIFsR3j8GSHl5vT/SP0EyDqM1RJe/LKwtJX/Syz/OqeDKnlTQP3bShW0VHbO/YgTvsg0hiz/ACw5kO1xwP9M83/DobKg/afqpbJzJ4z9a0V1ltCToPw9bmPiy3ek/aXFXhyn45T+z8Wdki77iP8+jtB59Qds/a8G0A4A11z/ah6bosofTP6Q/Iy1HnLs/dgHibVO1wD+WQ2rMbNDNv3ehccLvDMK/qZ3+57vqub8FA07TpHixv8tI3S241uU/O56aGzPW6j/1XiQJoW7sPzP31MeOOOg/awi7hpDl5D8hkAJf7OPeP1Yx9IpTKdo/9W3+i61F1j/jWftPbAvDv4k0eYAn/6e//DeZxNy8or8ZmARBFbDQvzyWam1k/sK/C21Cuebe7T8AAAAAAADwP9DFe7wTVug/EOYNQeKJtz/HXoIav9DqP0D02VbmP+0/JzgxPjlM6z+XkG8t7y3nP4DfwNIbb+s/VpLu4/Fw4T++LKZ3dTTtP7lpkc3GWe0/mdwJV0tO0L/uRoaUpcXBv1bVRat4s44/drWY0WDy6D9QPFOvXLrlPzXDX88GPLa/dZ++i14Ewr9M0uIgxmHmP/DOqbRnOd8/1IgPwj+X4j/nFsn0l9TXP/dFSnA3LZS/Qy5d8bgz0z9LQqI1DYnWP9uj8I4DV7u/f/HeIF/Glj9ycesPU1BTPzinJUJ9g7A/mOyArDvs0D83K2/SGmzgPy3DEWmt3uk/JQMpfggS6T/tgUaBHlLlP3x6CHPRKt4/cK5YLR/e2T/m5mq9CrHVP2aKuzZ0k8I/DK/FVUCxxz9Vsoyvbc3Gv11qCkkDUK+/lvtCaD8ohr9FeJoOE9CXP1jdSava89E/AcWrigP64D/2tvAMLlrpP17dim+HX+g/Q3f7CKHo5D+/n4YHYJDdP3qSc1bFztg/ZllIiqF01D8Gg0zifQDQv+X25FFt+qa/s59sHpXoqL/1iU/yWO3Uv2CbtvxVm8W/i4BIWiDc4T/QxXu8E1boPwAAAAAAAPA/GeTBghTBfT8HY2ia3LzcPzWOw8jKlOM/5ERBX6ac3j9suG1s4mzoP7Jln7JaVuw/eVUZtpEo4z8c4kFBtkHoP9Yo4fMwu+I/0tTFYN1yx79+yxzxYnirv5VJpxNV5o8/GeFCx65B5T9uH4RGZJziP8V56XdmYJe/5hityg62vL/zgHMGXr/jP7wBLB26Ud8/0JrVTFi53z/2EG84txXeP7xEevtBap4/L0CjSYZf1j8xjeKaXzjVP+kLI6pyhZm/fzwvOir2gj/1uUC56ZOiv1udfrbYIrY/U/N7J8y2bT9N9YmpwPxlP8Zx9OFZmUy/Gq6TAB12oD+lforngXiiPzBCivHRUoE/VJN6xJubZr+/kMd7fLlvv++nHIfsdYc/hvLuX1y4kz9akBx0Aj6hvwU2WvbjELa/Z6Usrq4Ju78cc0Yx02W8vx829I+GXp8//Bo8PFikpz9fIeRPVIq0P0eg4qAfzL0/E55Fc3qouz/CwX0BUBmtP1JRUJjRi6Q/tGNhSG3fpz8ZeekyRK+ov4zCex0dYZY/LLk7+BU3kz+TkkxmmXbHP6LyouKWyrU/Ywzb4gUprT8Q5g1B4om3PxnkwYIUwX0/AAAAAAAA8D9HM3Jh6OqQP4u9yKaYVIY/SeoMEs89mD9UxoUGYyCwP5UzA/FPDpU/R6lt6YoHmT+2I2ntTYV7P4OOFwMBs48/gxhALYJ+mr/V/4ccjBfRvyQlTqtQ0Ko/aJ7AQO8joj+tIQi/SfS0P4eW0w/mN6A/JBGcDWetwT+xLvHHCvmrP5eh1xWMJrQ/BaJvESVvsz8dxzdlPBy1v6Hdc2EHP4e/JdXzybq2rz/jFkvRKfK+P/NIhkwKkby/EVfLnZk4lb9g0q0FJql8vwg3yN6PJpk/MubnnE8E6T8UYfUv2/HpPxcyyqsjjeI/um/91Bhh2T8ZC8bIh2LVP1R0KfCeBtA/9Bvub83iyj9mRzrYKzTHPwl7mmry66M/cVaxHsypoj9HOHtEuYS4v+VEyAViJa2/Mug8NHKApb8flWdu5Yafv40p/ON4gOo/eAv9cNl36z9hb9wm0jDkP7mFEJxVXtw/cMuCZ+gZ2D+uM3EuGkbSP+cj29JXis4/e97wIUaMyj+VNOj0Akp+v0iQUR/JBKO/mSQRbdgtl78Xa4kNl9m5v+GZqH5w6qu/+AlbHYD+7T/HXoIav9DqPwdjaJrcvNw/RzNyYejqkD8AAAAAAADwP0Z9fP8bCe4/x5lVUUPc7T/lNyFw9sfcPzlUgVtuGeM/XTYjEIHA1D+D9XU5r5rpPyusPnSoMu4/E4+XSvv5wL+fgXp046y4v0eGxjVaNI4/mt8SKbQ74j8JPv6dYRHePzQ21PZ/5b6/02SliSEJwb+dO3CffTLgPyYe7VlaodI/+xhgR/c02T+E44iRuhrOP3Ezbf2k7pe/wgJlRkgmxj9dj15KiOnJP4aiX/LVoKy/C4k4Q88tgT+xJsoAmGJwP7m7NA/br58//kZRY6h55j96dq5aUuXpP69bZOIQe+c/ui1dFdm73z/99WUBIUXaPwGl0binbdM/FTzig3Fh0D//kpUjQMXLP6023qEPiKU/zzKOMnmRpD/VbDc2oKHBvyS+ig0XF7O/pXaDuviIqr/64kDJ4Xihv9Jl0n5b/ec/rdr0uz+p6z80uIHGCffoP7nNircci+E/EeRtqWJ13T9VOdOHnwPWP5Fqvdbtb9I/a8AeUD5uzz8RyrIw3qOGv3Wa2wWWfKS/fmx2p8qsmL8z7vJYaQbGv+2goeu827S/cgduolD87T9A9NlW5j/tPzWOw8jKlOM/i73IpphUhj9GfXz/GwnuPwAAAAAAAPA/1WN+cZTU7D8+WdnEiI/jP5IMUCV02+Y/Ux8OZr0y2T9dGvzW1MvrP31g0yaFle8/zqTzPdK0xb+5SapNQ66+v7dMJb4BMYk/DK4EofGu5D8nnukXnjzhP+Y7TAsgSru/1Lq0l949wr/ylLrsuJLiPzst+xZsEtY/uczgt//93D+oYCKNoiXSP66crIl2+Jm//JRqPT7wyj+ydPKL/0rPPwjTtz52T7G/TgecNfwPgj+7Av8XjnxQP5fScI9Dr6M/btsSmiSa5z8+WNVO++3pP3wPQ+DsMOM/y44myZ+Q2j9wrjnC2mHWP1C/YRaCUdE/feHjNAK+zj8gm5X0+fHKP+K16Sr+PbA/NuyHdhp5qT/OEZNI7uS3v7VgEKjZVa6/hoXwhmQ4qL9xZRQ5u6yhvyhvkkw9Uuk/0zD4lpHN6z+XV+tUmt/kP8xVt5HmcN0/FOn6kEgH2T+Jt4u6CWjTPyEozKC7/dA/XgHFoUfyzT9W/14PJ7yEvz4iYodeEau/1WOJra8qqL/N/Rkk1aa3v9z4oJVyc6e/LCLtfOYz7j8nODE+OUzrP+REQV+mnN4/SeoMEs89mD/HmVVRQ9ztP9VjfnGU1Ow/AAAAAAAA8D9oXHnCdJHcP+M79NI9teM/wYdff2uA1j9KimSzE/TpP91jYuXPzu0/aA5epRqbwb8nP1GLYjzBvymp5EVUzJA/UGvchSUD4z/+B+uaM8PfP7W73hsaFcG/WL6UIJM2wr8P+8Sgb5XgP02op3vMNtQ/7uhgClsc2z8W3Mf6HqbQP5WkMZQJkqC/eaGzmnbwyD/T1JArNJLMP2XBzuh1jK6/ATSkgI1olT8aloFDeaOQP4eZYb4ynqE/wbona7GPzT9QbZl1ZN3bP4ljwa5HXOg/GTASBAkW4z/8R7m6z+3eP82C15e5BNY/RwLIB+EE0z+AL3bBLtTPP4GtAKqG3LA/OOg43kGEsT/W3oSicjbOv6MGycoaFby/KrlNHv9JtL8MG7SRR5qrvz1YSM5fmdE/Y/NwJxHt3z8zeTb9C37qP351PA/5A+U/0vXw/xlT4T9AsCKNzzTZP/l/sJbAkNU/jDnCcbw30j8Rsvf+OAazv5UwfGJqaqC/E6mR2LvZnb+iKf8qynbWv12A/f+1iru/KPwYyWFU4T+XkG8t7y3nP2y4bWzibOg/VMaFBmMgsD/lNyFw9sfcPz5Z2cSIj+M/aFx5wnSR3D8AAAAAAADwP/KK8SEQPOc/1N+1uDFv3D9Od9n5p4jkP2JloFilgeI/Ko68LkqWzL8kRiZXmw2nv/r+FYC3eFa/GlohtpCm4z+pt+eqswXhPwL2l6okE4+/P49gzZJ5u78KBCb2bzPiPyiFv0A6ONo/dgJ2u4hF3T8OhieQwb7TPzLhEQaqxJW/QvbsllbAzz9devtbfLzSP3h5cB0rB7e/4zlbCKqtgT9yxXbqlBx6v8ST4YKEIaU/QLYEIfBJ3T9Oa2vMhC7kP3gqqWXuXes//wjPwBCC7D/KXTML0QXnP8HccVEaZN8/s3HtSJMd2j8O8Cz2VCvVP0wEe1BVQ8A/OiVwWks6yD/7Puqk9zu/v2g760K3G06/NJc2I8Khpz/jDTYy1QO2P3fic+1n4tw/XCfh64na4z8MwhQQOd7pPz3ayBBsbuo/lq6N63Nx5T/jdrP5dzrdP0nPBuvrotc/mYVLF3Cc0j9r7z+hX9fRv7KktvqQhJ6/eUdK3o9lm7/FhJmE/yvVvzfv6LPLbcW/owkXJRFn5j+A38DSG2/rP7Jln7JaVuw/lTMD8U8OlT85VIFbbhnjP5IMUCV02+Y/4zv00j214z/yivEhEDznPwAAAAAAAPA/P/5HdzIr5D8wPj1XwPnsPz/NhIltgOY/W++HYFPqwL9UvYiwMfmuv7Q0utmC2pQ/14S4m5jH5j/l2yuG6D/jP+xYiCyxd6e/doNhYfqswL9iurGoxnnlPy16BfMmsN8/flqSYwlK3z/fHYZIXqPgP4WEJU7YCne/XQ67k+Nj1z+Rwlf0QYDUP7yMTK51g5k/O41GxZC+pT9Wo9aoAvyMP36bv8yDwrg/UccThgP0zT8oO/LjSVvUP8plxQvBcd4//x3CzFwQ5T+jS0+27cjoP0JeyE8nZu0/e2ABUMXA6z8PQ2OdJFrnP2sCUOLPxOM/oX8aYVyr6D/1rrrJWbLIv1oy5mFc48S/j86qzbL3ur++bdh2LCl0P5WVwCcRJdA/b4pLgLH/1T+QjZAMgsreP+Fb69EUduQ/sHAqmYWQ5z9tWD0SFdTqP/GpTlL04ug/OmsnTIDR5D+oIXA6Fm/iv0B6fysKGuC/uY+HZ0Rzyb8fPPMb6qm9v/32j3ST2My/N9aentG82T9Wku7j8XDhP3lVGbaRKOM/R6lt6YoHmT9dNiMQgcDUP1MfDma9Mtk/wYdff2uA1j/U37W4MW/cPz/+R3cyK+Q/AAAAAAAA8D/YiHOHbvXgPyWqHlS7stg/UjNYvQbKyb+tVyHaq1c8v2dJGSf9/6a/6/DKKAkf4z/7UouClHjmP0b2jYbQRom/luSbpxyVoL/AvycYx6zfP4In66HyOuQ/0FPJKmz+5T/ObcOo/gDoP3A+m5w+17g/ittr+QSH5T9F2cIb56riPygymW3mKGQ/Lyn5eJDBuD/GqVAt2B+vP+3Dbr7QAck/TKF9AstZ5T8sL2fnyxLqP8j1z+y0lOs/1WqDjRJc5j/FkRtO5qPiP+T3mWcscto/c9QBWyLq1T+bTKI8zerRPxSnim2f9bI/26BbkzRiuz9KVPeZrY+rvyhv1qVhv6A/CHsb2J9jsD/Pedu3LWa3P4vTHuENjuQ/zlQou0n66D/NZlH07NnpP4Gybzit++Q/XCdrjylz4T++dN8nK3bYP+RjJronlNM/32ZrdHzwzj/91BgVWCzAv6qizJpFz6a/zbN/9rFlmL8dEI0Ip4LNv9V11NE//bu/CD/KG4fN6z++LKZ3dTTtPxziQUG2Qeg/tiNp7U2Fez+D9XU5r5rpP10a/NbUy+s/SopksxP06T9Od9n5p4jkPzA+PVfA+ew/2Ihzh2714D8AAAAAAADwPyWE+oP25es/bOIJftJDtb+SEvY1LQO8vyR5CLW6I54/OUdxKUYN5T+7iH9+k2LhPxb5pxb1NLi/IUefBRXewr8TTXqQ/wvkP6ZHYkXHbto/q8srrVD02z/tsBj/mD7dP0EXjfpuR4q/4SDwrxC70z8VBOKHBwHRP35BrXQXYaQ/FeaHD+i+nT9tAjz7BFaHP2UU8gWRqLQ/pre3SokJ5z8T5TuMhS7qP+imfdbzzuY/Y0LiNPb+3j/e642ZTsPZP1wLxFR68tI/vpIDf7sT0D/hZm2WREDLP+tGf4cfTaU/munKSu+Voz9jI1Zk+SbBv2A5fiqKSbO/6Wx0KGBzq78IpxeAPQmjvyWVMHkMoug/h1ay9bv+6z9MZ2fdvGfoPxBt+ZzzI+E/2WOPrbDk3D81PKNoIoTVP5WPv8AZHNI/Yt6DpXv7zj8u61cvKPWHv8iDKgRWD6O/K88Tqk2ymr8TBgYJrJfEvw7J85T5ZbS/UdBRD1qw7j+5aZHNxlntP9Yo4fMwu+I/g44XAwGzjz8rrD50qDLuP31g0yaFle8/3WNi5c/O7T9iZaBYpYHiPz/NhIltgOY/JaoeVLuy2D8lhPqD9uXrPwAAAAAAAPA/xmVQm2Nsxb9kp6toYlS/v3t8ytM9Uos/3mECZqum5D+9nS8fai7hP8DQJ6gnery/JkdTIR04wr9hIGh1D3niP4Ja0vft1dU/0LjKRybj3D8SnzMJQ7TRP0WSt6yZsZi/e8SV2fA8yj86tJk0C5TOPxNk04x8X7G/CTkdPnl9gz9RWFY2XWZmPxNfDq/R6KQ/5Avf51HKd79jZoAIRF+fv0z9RIAIq7m/9L8C79i5wr9q8eSYn2/Dv3ZgaQVRdsa/lEewUSEtx78rvzvkn9bFv37GtRQBl8C/rmTCZ80Zv7/zWHZuhMfuP1k658sDNO4/f/OhOLfM7D+LQtzymTfrP1fbwFAZ7bi/bUm8zYmQw7/7SAXtNRrOv2y47pMhYtO/UpDJ7XAk1b9rU63ibpzXv8usQYE4pti/zrUXJ8bK2L+4UxeLk0+1P75EALSR/aY/CUcAFK2RkT/qA7AY0m/DP8VQ0y/9HMs/hHA652qaxr+Z3AlXS07Qv9LUxWDdcse/gxhALYJ+mr8Tj5dK+/nAv86k8z3StMW/aA5epRqbwb8qjrwuSpbMv1vvh2BT6sC/UjNYvQbKyb9s4gl+0kO1v8ZlUJtjbMW/AAAAAAAA8D844Fe4sRuVPxHs11ikL5o/92DE9cXp47/0mC/yqe7kv6D+lf7nIZc/+IW8gRwosT9OozjxKqLOv9UPGuMY98e/0omuC6OP5b96Qt4Wii3RPyYTFvZ1ubG/UB8TsNFEcb/+aSCHzMLIv5pu3VYzmug/yfVDPehmLT8zI0DVwGeoP5tvKvR4pqY/sajJGC57s7+TyGjEKXq+v7R+brv/CrS/LCeD4Ft7fb8Gsk1Yon1uP+SLVuQMD4E/kGn/4HOpbT9yvoPzEHhFP6EW8LT7hIk/jNlYkQS7oD84OtIRKyGVv6mbf5677Wo/Lt7R0r0llT9QZMn21h6cP0t/oX2N+bu/5E6Mcd+Yxr9zQJ4x9vO4v3kfMxk6fZS/8YbWi6Uhf784n+gbF7Vuv2WbLu/fnna/fQLWNeuogr9kJGe5v9Krv8pj9Pj8FVW/07EOqRPHgj/ZHdKvTpiyvwHvnK5vFay/W/GT8mO6xL/uRoaUpcXBv37LHPFieKu/1f+HHIwX0b+fgXp046y4v7lJqk1Drr6/Jz9Ri2I8wb8kRiZXmw2nv1S9iLAx+a6/rVch2qtXPL+SEvY1LQO8v2Snq2hiVL+/OOBXuLEblT8AAAAAAADwP8SKFhDjmaq/clxRTxcyrL/XLajkiamgv6NaxGvzY7e/NdEyxLRqu792ZIK4cs2wvycw/4g6XKe/rXTOycsYl79Ijqd77ZebPwROw3Vu1p2/tR2idcFSfL9IInOvZdSRv4Eex1RtRIG/ryU124fMrT/7lUxE14axP258hbB5eU8/QJMknTQAkz/OsA5NZbeYP1+zczxWlaA/OyCMbehDfj8txFBaWP6Qv8nGIhNZVa6//w845Z/fnL+SohDL1dqMv9Em0GcuxpK/N8Z/+sfPsr9b3TPI7KmhPy9BwEPcfKM/lN2ASFmqcb/fJS4gQFeiv5+65MknrXc/uNYgF9VCij9kaKXfZVCWPwXna79KQHa/BdONrpvjn7/3OhVEVo6wvzALDNQNiqG/Vwm8XO13kr+yzOB8j0GkP9sDOn1sL7E/hQtK4rv2p79vzO7/0MlzP504g9mBGL8/c27wUWkVlT9W1UWreLOOP5VJpxNV5o8/JCVOq1DQqj9HhsY1WjSOP7dMJb4BMYk/KankRVTMkD/6/hWAt3hWv7Q0utmC2pQ/Z0kZJ/3/pr8keQi1uiOeP3t8ytM9Uos/EezXWKQvmj/EihYQ45mqvwAAAAAAAPA/r6ozDGstqr+ZQqoeJV+xv2JXxAIfCom//0GTzVPDkr/vHVWbl+GHv2GJl9/MKXA/H8I8Y0yXq7+1Y4HG1KWfv4k01qEUbZA/QTq08W0+mL+qBBhxfsymvzxoXUvdCqQ/FqrGkOgnub8OZxy314etv9k9vp8Ppqm/63S+mjKL2D+TaYdsmTbeP0YCzLLN7+I/LC7V7Kct5T/XvC17HT/kP5oAf+zuDd8/xmwhLTdI2z+PbTvKFSLXP6KKj7e0Bcs/xY4QhScF0T9A/A2bTWPjv3+2OyfoD+G/FzxxP2ID3r/ORezCZg/av+4xaAbfVd0/Txwlatcv4j+JpjytZ9vlPzwFe23s0uc/Z8dwjCH+5j97xi1i6tbiP31HTDBPBeE/NBVLS9kz3j9NB1IVFwfUv3qr0pElb7G/ySXmU2jLo79KlQQJJMrPv+WcwCzaDtG/Mqe//S0f5T92tZjRYPLoPxnhQseuQeU/aJ7AQO8joj+a3xIptDviPwyuBKHxruQ/UGvchSUD4z8aWiG2kKbjP9eEuJuYx+Y/6/DKKAkf4z85R3EpRg3lP95hAmarpuQ/92DE9cXp479yXFFPFzKsv6+qMwxrLaq/AAAAAAAA8D+LFeFMY1HtP3DjS6zPEJi/np8LoR1Qtb+JpwDUdRnkP7FiBFlJDOA/nrRi82hl6j/v6AkeXYnMP8kKdoLSgIw/5lOEE3DI0j8w8TxtVDLZP9hQ5HFo6tq/mwaBhKm2rz/lEtGjUPaQPwCvrTRilak/Ots2HUct0z+J6j2OP9PXP5Qi8SPbHd8/kIc0ytOQ4j/RoMqeqsnjP+o1m+pgJOQ/XK5S4+Bz4T+AIVg6mbzdP9gmQnzBBNY/K5ZbyaIa3T92EsFVZVzkvw73aoWHNOO/PsZOKuRD4b+TfJhB2CHdvxL2djUrWdg/eXZyWwp43j95Nlbpo9fiP8pr5e2j7uU/u6vWtqpJ5z/6wD/5J0nnP44UXHc44+Q/fJaEq0iQ4j9ZxRFXDlvYv9rS7tlN086/4Jv74O27rb8kRugBWbPEv2lh1mdsY9C/QQ3k0gXG4T9QPFOvXLrlP24fhEZknOI/rSEIv0n0tD8JPv6dYRHePyee6ReePOE//gfrmjPD3z+pt+eqswXhP+XbK4boP+M/+1KLgpR45j+7iH9+k2LhP72dLx9qLuE/9Jgv8qnu5L/XLajkiamgv5lCqh4lX7G/ixXhTGNR7T8AAAAAAADwP0BXUpVyRpi/uP9ZGYNHqr9wmbvrYyziPzn6+p81AOE/kdiyPnJk7T87w/EQFTrRP31A3MVDpac/xN1qvkby2T8PkWt0RxngP7yxj47829y/BroEV4r7sz++nyMhrQ+aP7shA9BXDrU/lsT/bPs+wL8sa5P2Lr2/v7Au+PAr1rC/B9feVE2Whb+/7ZEuz8qFv6IaVM4ZaoS/b0gG7GlAh7+D3vXzsGg4P6klNYfHdpY/ENKnnWfskj8vj3ps4yiQv0NMrl4WXWy/l3+cef3vYz8WQh7+nq9+Pw81Bl9+6r+/Nt+NYafgvr9hIQlAB4muv9o9eNH2m4a/hjTaBAZUj78U8VxNRT+Jv68xJxlT3Ye/k46QdoUtRr8w5JSbSm2fv+UZ0nPCf1g/omjuIEWvi7+Gb2bwS7qgv2APNhHlJJy/g5yXHeQyv781w1/PBjy2v8V56XdmYJe/h5bTD+Y3oD80NtT2f+W+v+Y7TAsgSru/tbveGxoVwb8C9peqJBOPv+xYiCyxd6e/RvaNhtBGib8W+acW9TS4v8DQJ6gnery/oP6V/uchlz+jWsRr82O3v2JXxAIfCom/cONLrM8QmL9AV1KVckaYvwAAAAAAAPA/Q0CEq3Nj7D/pg6mUdLOhv375QGKHf0O/hLo4XSCUl78n7mNGvXWDv0HItAdxiog/49cgvGFJfz/iftSTvp11P7vsXm+INY4/A+LHazrXSr+CDCGBdyOCv5GDw/BkCYm/HZPe4o3mvb9yFjKBVnLBv5iealNBZ8K/4gB6IJoFuL/c1KdOvRyovwNb+LoqAJC/0W+ffXVWjb+GOt+sTd9zP4734QBAX6o/J0w87njTqD96Dl3kECCrP5FsBUEm0ZK/LH8rnmSQk79x0TIA3WqCv1PhFB2IPb6/bMQJdV+4wb/PsYiTY3HBv2RpnIhNLLa/jigbGQBzpr/H/c/Y0qaOvwZAjtdS5oe/sIX4zr9Kej+d5X7iULmlvzu+XB5a6pi/jhtvD0ADlb8U0xeYrnvJPy2NaSmWJG6/Cc5LbVm3wr91n76LXgTCv+YYrcoOtry/JBGcDWetwT/TZKWJIQnBv9S6tJfePcK/WL6UIJM2wr8/j2DNknm7v3aDYWH6rMC/luSbpxyVoL8hR58FFd7CvyZHUyEdOMK/+IW8gRwosT810TLEtGq7v/9Bk81Tw5K/np8LoR1Qtb+4/1kZg0eqv0NAhKtzY+w/AAAAAAAA8D9GeGkYaiCzvyPqt/DSa5C/uGVhtFTgo78xK6Mf33Siv37yNNQ5KJg/8kCFwE5gkT96v1PjBgOSPy6JJFoKyIU/zsqIZa95kT+DAQxYEO5pP1y4QM2Rl2A/f4FPGRR11j+upCB406zdP2iMb/bipOI/kIlD15+54z8uRhximt3hP3/zjOiOddk/AdcphHoI1T+JCB+LDiPRP1PBXiuwN78/SbEFOC7Dxz/0owrWgGPMv5jmbjn2esO/TbAM6vuvvL8sy765nnWzv8u7U8lqs9o/p1Gfx+UO4T/Puypz25nkPwzhwRAEWuU/p7k5Wadj4z/5tK+c8r3cP4lIzRO/6Nc/vIIN3Cre0z/KzY/cqFDQv+tqx84LDqO/QfT3wOtLk7890QC/aJHMvwny0B76l8S/8HMeZvgC4z9M0uIgxmHmP/OAcwZev+M/sS7xxwr5qz+dO3CffTLgP/KUuuy4kuI/D/vEoG+V4D8KBCb2bzPiP2K6sajGeeU/wL8nGMes3z8TTXqQ/wvkP2EgaHUPeeI/TqM48Sqizr92ZIK4cs2wv+8dVZuX4Ye/iacA1HUZ5D9wmbvrYyziP+mDqZR0s6G/RnhpGGogs78AAAAAAADwP2A+yccu++Q/axUiPOEG3z9cYCse7ofVP0bfl7nE1JS/6uXFyCYu0T/ah8TDf/LTP3IQc9wFSrq/QlMRnucenT8ikxB/QRiCP9UUCL+fybU/G5z+uPjsxj+jXIZ1jjbPPxJQKI2hhNc/ucIs0Qnb4D9b8sKk1szlPxJ94+obw+E/Ou9BUKBG4D+tFD7RYmPdP6t22VjbL9g/zSGU8Z/p3D/oxzrMMyXHv7WCgBIRzsO/TJ8G1dovu7/VChmihgWwv25vZfl7zs4/QfnwL9Ql1D/krC7C0rjbP7s73X9aveI/kaSH4EFw5z98aj9yT/viP5v07uzDOeE/TN2utLTq3j9vl44hfBviv9yKJ3kxBsK/JyObaGP1wr9zoCaXQxa7v+rDRnVHE8m/KJ4C6DFl1z/wzqm0ZznfP7wBLB26Ud8/l6HXFYwmtD8mHu1ZWqHSPzst+xZsEtY/Taine8w21D8ohb9AOjjaPy16BfMmsN8/gifrofI65D+mR2JFx27aP4Ja0vft1dU/1Q8a4xj3x78nMP+IOlynv2GJl9/MKXA/sWIEWUkM4D85+vqfNQDhP375QGKHf0O/I+q38NJrkL9gPsnHLvvkPwAAAAAAAPA/KZWQWxYg4D9ZR68XbqbeP2PM/2c8yoa/tONNK46t2j8KLck+uzrdP9zc+rxanbO/7mw+bKaRnD/dBo/vSt+SP3vyFz3x4sA/etRT6CkRzj/yUA4h48rSP21pHECjAdk/f6YlQH6U3j8YefZk6rbgP7HIGNqOZOM/nmtjbgds5D8qF5S7TaXjPwyLFS1RhOA/EkzBUNhs3j8VfIKYk/nkv+44c0KafeS/QLO+ju9K479oGc96edfgvz0aYq0VgtQ/n0NFepDS2T+o5z5uACTgP6pRAnNHJOM/Whlk3Kjm5D8qU9F2m2HnP4PaHjkBBOg/FiRlOOL85j84fK79VI/VvwK1DD7ldNO/SKJbUs8gzr9GhPaAxay7v8OeoLK77sW/wTReIFsR3j/UiA/CP5fiP9Ca1UxYud8/BaJvESVvsz/7GGBH9zTZP7nM4Lf//dw/7uhgClsc2z92Ana7iEXdP35akmMJSt8/0FPJKmz+5T+ryyutUPTbP9C4ykcm49w/0omuC6OP5b+tdM7JyxiXvx/CPGNMl6u/nrRi82hl6j+R2LI+cmTtP4S6OF0glJe/uGVhtFTgo79rFSI84QbfPymVkFsWIOA/AAAAAAAA8D/2biWax0/QPy6WTNkBGKY/a96u/13a2j9EuVfRLP7gP/BhArM9EN6/vtLKVwhEsD+MyW2gs+6OPxD1eUmgYbU/Cyfh+081zD/X8D8flNLSP7HyQ/Rcddo/B6dIALLx4D9CjKcJ3VzjPycHcA0tROY/P4sng0lD5j+6SKetEibkPwYAEUNYceE/73KZxWT84j8WsdMBF53QP3L0UF9RvNM/DBgDz3iA1z/iwH8s03jePxyM5G73m8c/jmE6wYNlzz8+tj04xqrVPyZlyb14eds/+1wCVvzK3j8WkU4lc1vhPyzmRTsOzOA/BuAYh65l3D91rq5bOufcv3yyEv0DrNq/o4MlZVWi0b+EgnbRaljBv5Os6QXtU8e/Bkh5eb0/0j/nFsn0l9TXP/YQbzi3Fd4/Hcc3ZTwctb+E44iRuhrOP6hgIo2iJdI/FtzH+h6m0D8OhieQwb7TP98dhkheo+A/zm3DqP4A6D/tsBj/mD7dPxKfMwlDtNE/ekLeFoot0T9Ijqd77ZebP7VjgcbUpZ+/7+gJHl2JzD87w/EQFTrRPyfuY0a9dYO/MSujH990or9cYCse7ofVP1lHrxdupt4/9m4lmsdP0D8AAAAAAADwP74NmjQ2Tp8/agPtYyGF4D9zrT02R8LWP0puCCxXKtk/6uEXkuNgwD9qApmOS+S7P2uJKUfK6s8/skvf8w25ar+7YOFP+AWQv1W+aT4a74i/FGyxj3KkcT9iwaJHrbKhPyRlheODX70/CWTQlXqOuD+Vz3VzgvW0P5k5gcQHiLY/18Z3CInhvT8wmgTKDdWuv72iftSANrW/b4/dUdoFsr8B0N/EfjOwvzTvEpDc6oK/dEYwpVxPjr+DhC/dN/aVv5bX8xgPXI+/SIPx0eYsiD+n0XpohMmzP/bxfyBeRLA/1Js996lMsD8MJLmJZtOlv2O86BmoHrW/Pi6MO5/7kL8567rDABiiP/HdchKiwLG/QTIOozVEl7/3RUpwNy2Uv7xEevtBap4/od1zYQc/h79xM239pO6Xv66crIl2+Jm/laQxlAmSoL8y4REGqsSVv4WEJU7YCne/cD6bnD7XuD9BF436bkeKv0WSt6yZsZi/JhMW9nW5sb8ETsN1btadv4k01qEUbZA/yQp2gtKAjD99QNzFQ6WnP0HItAdxiog/fvI01DkomD9G35e5xNSUv2PM/2c8yoa/LpZM2QEYpj++DZo0Nk6fPwAAAAAAAPA//2OMVutvOb9+l5pWlrhdv8XSwAMRZIA/xJ99FJSa2L/xvCf6IxTev8h3yK/LI9E/BR9FO3CXwT8377L2BTTIPwFXYHdAW9E/ex5eMIaS2D9cpGcN34TeP0lTBslOBuc/OXtfAf574j+DVheZz5ngPzSLrlGHst0/HodOsgTt5D8X9D5zHrGqvgEQvPOZyXK/9lUxvKTPPz/CElWJvJSXP3OBwRbyvsE/rAMrwbhGyD/UZF8qQ9TQP54hHt4uDNc/gSsVGEEW3D/nF7/DZBDkP7NfZZSV5N8/KNyfCcZh3D/YuIVTMN3Yv1+qe3cPHOC/380Dg985uL9yo/7z4KJcvzA/onYlRZa/LKwtJX/Syz9DLl3xuDPTPy9Ao0mGX9Y/JdXzybq2rz/CAmVGSCbGP/yUaj0+8Mo/eaGzmnbwyD9C9uyWVsDPP10Ou5PjY9c/ittr+QSH5T/hIPCvELvTP3vEldnwPMo/UB8TsNFEcb+1HaJ1wVJ8v0E6tPFtPpi/5lOEE3DI0j/E3Wq+RvLZP+PXILxhSX8/8kCFwE5gkT/q5cXIJi7RP7TjTSuOrdo/a96u/13a2j9qA+1jIYXgP/9jjFbrbzm/AAAAAAAA8D/BAAd/8mrqPy7BYbSbVHi/viVxdsbtc794JHfJStB5v4EDZ3kaunS//cz3W6/XvT9KYj7G6/TEP8b+JeqqSM4/dsr+IwLD1T9OaeOEmcDaP9fgHkfRveM/n/vvmWVl4D/3FdnoCNndP4bPpf6d49o/LT7If/Eu4j/7UOoS/CDIvxNwlW55v8m/Df0LoFwSyr+9zKHfwPbFv+okJZtmDMU/xsnYi1upzD+5XIEqWcbTPwCPpRc7N9s/7H08okqO4D+SRAzuSS3nP1KXUsEPMOM/0CL0ig9r4T8g4GVInB7Wv/4cxAzIjty/GQh5Hoz0tL+7b0fsPtpPP46854UGQJe/zqngyp5U0D9LQqI1DYnWPzGN4ppfONU/4xZL0Snyvj9dj15KiOnJP7J08ov/Ss8/09SQKzSSzD9devtbfLzSP5HCV/RBgNQ/RdnCG+eq4j8VBOKHBwHRPzq0mTQLlM4//mkgh8zCyL9IInOvZdSRv6oEGHF+zKa/MPE8bVQy2T8PkWt0RxngP+J+1JO+nXU/er9T4wYDkj/ah8TDf/LTPwotyT67Ot0/RLlX0Sz+4D9zrT02R8LWP36XmlaWuF2/wQAHf/Jq6j8AAAAAAADwP5JKzxTSM86/YRaa9t6Ugr//CnqHAlmFv8bdpu3eW3C/yaeXZp7tnz8Ci+7VpTWjP85gTvgtV5o/x0xRT9SGkD9pJwquCfmNP7mkSXjiv3U/5yy30TcYfT8MSmME0WZyPyBgOUizdoQ/GY54osHqiD8X9PEKckjoP/7kFQj5n+k/qsDCEi0j6T/JZoOUzZPoP+NIUkHbCai/OFTbNsjor7831x3Bf+23v+2Ss6F/AMG/Peitv0HFxL9I7Idg+KzIv+/ppezyH8q/K2hGKOICzL+7zG6vHxyFv6enIZc6A6O/DWLy/nvCp7++DMp2ZbCCv/5+XXU4arg/dtKFbRUds7/bo/COA1e7v+kLI6pyhZm/80iGTAqRvL+Gol/y1aCsvwjTtz52T7G/ZcHO6HWMrr94eXAdKwe3v7yMTK51g5k/KDKZbeYoZD9+Qa10F2GkPxNk04x8X7G/mm7dVjOa6D+BHsdUbUSBvzxoXUvdCqQ/2FDkcWjq2r+8sY+O/Nvcv7vsXm+INY4/LokkWgrIhT9yEHPcBUq6v9zc+rxanbO/8GECsz0Q3r9KbggsVyrZP8XSwAMRZIA/LsFhtJtUeL+SSs8U0jPOvwAAAAAAAPA/2v0x+VBFVD8lNVH8Z/Bmv0+ap48ZmWY/VcR737u4ez8Lq3Lk2JZ2P1JlrTOfLJI/rGgFvQ08qT9vRPYjupGzPyCeUuM8YLk/Mpkmi+lUtz945j4+PoO5P5CjNZ8/4rg/SGMbM48ZuT8NdeMEiCdkP2tj8iXAU28/HIf6pJ0FmT+WPCyJQPS0P/3aB+zynnU/UGcLj1p7Q7+nYyLQ90eFP8FllwA/sqI/vQhjRGAKsD89xozHux22PwxfJKwCCrQ/2H38stogtD8Mpx61sN2wv0vHx4W3zrO/jbXp4aQLq79JNBqjZsp5v7yivoVp77C/YgTvsg0hiz9/8d4gX8aWP388Lzoq9oI/EVfLnZk4lb8LiThDzy2BP04HnDX8D4I/ATSkgI1olT/jOVsIqq2BPzuNRsWQvqU/Lyn5eJDBuD8V5ocP6L6dPwk5HT55fYM/yfVDPehmLT+vJTXbh8ytPxaqxpDoJ7m/mwaBhKm2rz8GugRXivuzPwPix2s610q/zsqIZa95kT9CUxGe5x6dP+5sPmymkZw/vtLKVwhEsD/q4ReS42DAP8SffRSUmti/viVxdsbtc79hFpr23pSCv9r9MflQRVQ/AAAAAAAA8D/7Wkko4X7qPwG4lP2eA6Q/zYQfKk2qYr85Yf2S7EYvvwcKpmnyqns/9YDJml48kj/a+WomQ/KfP1rDsNij5q4/vN9ocW0CtD9NEGH0ER25P5gnnSHtirk/zdAlP6UtsD9LJ1n2X2ikP8ASmg/wGaY/3nXSMn3MqT9PHSo16gS3Py1w7pMWG2a/Dy7+TuJPfb/v7WmejWhYv7LIZbjpBWc/QH2jGhLXiD/Ag8NPyUqkPx3h2soh3Kw/g7KcGk2Lsj/p9+lGeD2Yv52WRvQus7G/IviERru9uL/mxQmdOriQv6SgRlK4n6W/wAsOZDtccD9ycesPU1BTP/W5QLnpk6K/YNKtBSapfL+xJsoAmGJwP7sC/xeOfFA/GpaBQ3mjkD9yxXbqlBx6v1aj1qgC/Iw/xqlQLdgfrz9tAjz7BFaHP1FYVjZdZmY/MyNA1cBnqD/7lUxE14axPw5nHLfXh62/5RLRo1D2kD++nyMhrQ+aP4IMIYF3I4K/gwEMWBDuaT8ikxB/QRiCP90Gj+9K35I/jMltoLPujj9qApmOS+S7P/G8J/ojFN6/eCR3yUrQeb//CnqHAlmFvyU1Ufxn8Ga/+1pJKOF+6j8AAAAAAADwPy8uDShsFro/7UiSToPKnj+pi/iPZuyiP33EkHA8jLM/l0vLNKjAuj8jJYobnILBPxt+SnMKt8k/SUDjpMdhxj95N4EI9rbCP/HL6eFwHsE/IjQGnODFyD8JJlaIHyelP5zcAvsOka0/z8rSFyl8tT+3n3dqq1LBP1uRwN/3G5s/2J6CoRrKnD8m3M2EdGmvP5OKr44IpbQ/cFcNqGkguj+nrFkZLubDP1nKzVhM3ME/tZMSFMNwvT+ICxIp+fu8vyXgeC3hWcS/bAxlP5qVp78C7Cig+4qhvzYrRvpTHa6/0zzf8OhsqD84pyVCfYOwP1udfrbYIrY/CDfI3o8mmT+5uzQP26+fP5fScI9Dr6M/h5lhvjKeoT/Ek+GChCGlP36bv8yDwrg/7cNuvtAByT9lFPIFkai0PxNfDq/R6KQ/m28q9Himpj9ufIWweXlPP9k9vp8Ppqm/AK+tNGKVqT+7IQPQVw61P5GDw/BkCYm/XLhAzZGXYD/VFAi/n8m1P3vyFz3x4sA/EPV5SaBhtT9riSlHyurPP8h3yK/LI9E/gQNneRq6dL/G3abt3ltwv0+ap48ZmWY/AbiU/Z4DpD8vLg0obBa6PwAAAAAAAPA/",
|
|
"dtype": "f8",
|
|
"shape": "57, 57"
|
|
},
|
|
"zmax": 1,
|
|
"zmid": 0,
|
|
"zmin": -1
|
|
}
|
|
],
|
|
"layout": {
|
|
"height": 800,
|
|
"template": {
|
|
"data": {
|
|
"bar": [
|
|
{
|
|
"error_x": {
|
|
"color": "#2a3f5f"
|
|
},
|
|
"error_y": {
|
|
"color": "#2a3f5f"
|
|
},
|
|
"marker": {
|
|
"line": {
|
|
"color": "#E5ECF6",
|
|
"width": 0.5
|
|
},
|
|
"pattern": {
|
|
"fillmode": "overlay",
|
|
"size": 10,
|
|
"solidity": 0.2
|
|
}
|
|
},
|
|
"type": "bar"
|
|
}
|
|
],
|
|
"barpolar": [
|
|
{
|
|
"marker": {
|
|
"line": {
|
|
"color": "#E5ECF6",
|
|
"width": 0.5
|
|
},
|
|
"pattern": {
|
|
"fillmode": "overlay",
|
|
"size": 10,
|
|
"solidity": 0.2
|
|
}
|
|
},
|
|
"type": "barpolar"
|
|
}
|
|
],
|
|
"carpet": [
|
|
{
|
|
"aaxis": {
|
|
"endlinecolor": "#2a3f5f",
|
|
"gridcolor": "white",
|
|
"linecolor": "white",
|
|
"minorgridcolor": "white",
|
|
"startlinecolor": "#2a3f5f"
|
|
},
|
|
"baxis": {
|
|
"endlinecolor": "#2a3f5f",
|
|
"gridcolor": "white",
|
|
"linecolor": "white",
|
|
"minorgridcolor": "white",
|
|
"startlinecolor": "#2a3f5f"
|
|
},
|
|
"type": "carpet"
|
|
}
|
|
],
|
|
"choropleth": [
|
|
{
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
},
|
|
"type": "choropleth"
|
|
}
|
|
],
|
|
"contour": [
|
|
{
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
},
|
|
"colorscale": [
|
|
[
|
|
0.0,
|
|
"#0d0887"
|
|
],
|
|
[
|
|
0.1111111111111111,
|
|
"#46039f"
|
|
],
|
|
[
|
|
0.2222222222222222,
|
|
"#7201a8"
|
|
],
|
|
[
|
|
0.3333333333333333,
|
|
"#9c179e"
|
|
],
|
|
[
|
|
0.4444444444444444,
|
|
"#bd3786"
|
|
],
|
|
[
|
|
0.5555555555555556,
|
|
"#d8576b"
|
|
],
|
|
[
|
|
0.6666666666666666,
|
|
"#ed7953"
|
|
],
|
|
[
|
|
0.7777777777777778,
|
|
"#fb9f3a"
|
|
],
|
|
[
|
|
0.8888888888888888,
|
|
"#fdca26"
|
|
],
|
|
[
|
|
1.0,
|
|
"#f0f921"
|
|
]
|
|
],
|
|
"type": "contour"
|
|
}
|
|
],
|
|
"contourcarpet": [
|
|
{
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
},
|
|
"type": "contourcarpet"
|
|
}
|
|
],
|
|
"heatmap": [
|
|
{
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
},
|
|
"colorscale": [
|
|
[
|
|
0.0,
|
|
"#0d0887"
|
|
],
|
|
[
|
|
0.1111111111111111,
|
|
"#46039f"
|
|
],
|
|
[
|
|
0.2222222222222222,
|
|
"#7201a8"
|
|
],
|
|
[
|
|
0.3333333333333333,
|
|
"#9c179e"
|
|
],
|
|
[
|
|
0.4444444444444444,
|
|
"#bd3786"
|
|
],
|
|
[
|
|
0.5555555555555556,
|
|
"#d8576b"
|
|
],
|
|
[
|
|
0.6666666666666666,
|
|
"#ed7953"
|
|
],
|
|
[
|
|
0.7777777777777778,
|
|
"#fb9f3a"
|
|
],
|
|
[
|
|
0.8888888888888888,
|
|
"#fdca26"
|
|
],
|
|
[
|
|
1.0,
|
|
"#f0f921"
|
|
]
|
|
],
|
|
"type": "heatmap"
|
|
}
|
|
],
|
|
"histogram": [
|
|
{
|
|
"marker": {
|
|
"pattern": {
|
|
"fillmode": "overlay",
|
|
"size": 10,
|
|
"solidity": 0.2
|
|
}
|
|
},
|
|
"type": "histogram"
|
|
}
|
|
],
|
|
"histogram2d": [
|
|
{
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
},
|
|
"colorscale": [
|
|
[
|
|
0.0,
|
|
"#0d0887"
|
|
],
|
|
[
|
|
0.1111111111111111,
|
|
"#46039f"
|
|
],
|
|
[
|
|
0.2222222222222222,
|
|
"#7201a8"
|
|
],
|
|
[
|
|
0.3333333333333333,
|
|
"#9c179e"
|
|
],
|
|
[
|
|
0.4444444444444444,
|
|
"#bd3786"
|
|
],
|
|
[
|
|
0.5555555555555556,
|
|
"#d8576b"
|
|
],
|
|
[
|
|
0.6666666666666666,
|
|
"#ed7953"
|
|
],
|
|
[
|
|
0.7777777777777778,
|
|
"#fb9f3a"
|
|
],
|
|
[
|
|
0.8888888888888888,
|
|
"#fdca26"
|
|
],
|
|
[
|
|
1.0,
|
|
"#f0f921"
|
|
]
|
|
],
|
|
"type": "histogram2d"
|
|
}
|
|
],
|
|
"histogram2dcontour": [
|
|
{
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
},
|
|
"colorscale": [
|
|
[
|
|
0.0,
|
|
"#0d0887"
|
|
],
|
|
[
|
|
0.1111111111111111,
|
|
"#46039f"
|
|
],
|
|
[
|
|
0.2222222222222222,
|
|
"#7201a8"
|
|
],
|
|
[
|
|
0.3333333333333333,
|
|
"#9c179e"
|
|
],
|
|
[
|
|
0.4444444444444444,
|
|
"#bd3786"
|
|
],
|
|
[
|
|
0.5555555555555556,
|
|
"#d8576b"
|
|
],
|
|
[
|
|
0.6666666666666666,
|
|
"#ed7953"
|
|
],
|
|
[
|
|
0.7777777777777778,
|
|
"#fb9f3a"
|
|
],
|
|
[
|
|
0.8888888888888888,
|
|
"#fdca26"
|
|
],
|
|
[
|
|
1.0,
|
|
"#f0f921"
|
|
]
|
|
],
|
|
"type": "histogram2dcontour"
|
|
}
|
|
],
|
|
"mesh3d": [
|
|
{
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
},
|
|
"type": "mesh3d"
|
|
}
|
|
],
|
|
"parcoords": [
|
|
{
|
|
"line": {
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
}
|
|
},
|
|
"type": "parcoords"
|
|
}
|
|
],
|
|
"pie": [
|
|
{
|
|
"automargin": true,
|
|
"type": "pie"
|
|
}
|
|
],
|
|
"scatter": [
|
|
{
|
|
"fillpattern": {
|
|
"fillmode": "overlay",
|
|
"size": 10,
|
|
"solidity": 0.2
|
|
},
|
|
"type": "scatter"
|
|
}
|
|
],
|
|
"scatter3d": [
|
|
{
|
|
"line": {
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
}
|
|
},
|
|
"marker": {
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
}
|
|
},
|
|
"type": "scatter3d"
|
|
}
|
|
],
|
|
"scattercarpet": [
|
|
{
|
|
"marker": {
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
}
|
|
},
|
|
"type": "scattercarpet"
|
|
}
|
|
],
|
|
"scattergeo": [
|
|
{
|
|
"marker": {
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
}
|
|
},
|
|
"type": "scattergeo"
|
|
}
|
|
],
|
|
"scattergl": [
|
|
{
|
|
"marker": {
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
}
|
|
},
|
|
"type": "scattergl"
|
|
}
|
|
],
|
|
"scattermap": [
|
|
{
|
|
"marker": {
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
}
|
|
},
|
|
"type": "scattermap"
|
|
}
|
|
],
|
|
"scattermapbox": [
|
|
{
|
|
"marker": {
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
}
|
|
},
|
|
"type": "scattermapbox"
|
|
}
|
|
],
|
|
"scatterpolar": [
|
|
{
|
|
"marker": {
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
}
|
|
},
|
|
"type": "scatterpolar"
|
|
}
|
|
],
|
|
"scatterpolargl": [
|
|
{
|
|
"marker": {
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
}
|
|
},
|
|
"type": "scatterpolargl"
|
|
}
|
|
],
|
|
"scatterternary": [
|
|
{
|
|
"marker": {
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
}
|
|
},
|
|
"type": "scatterternary"
|
|
}
|
|
],
|
|
"surface": [
|
|
{
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
},
|
|
"colorscale": [
|
|
[
|
|
0.0,
|
|
"#0d0887"
|
|
],
|
|
[
|
|
0.1111111111111111,
|
|
"#46039f"
|
|
],
|
|
[
|
|
0.2222222222222222,
|
|
"#7201a8"
|
|
],
|
|
[
|
|
0.3333333333333333,
|
|
"#9c179e"
|
|
],
|
|
[
|
|
0.4444444444444444,
|
|
"#bd3786"
|
|
],
|
|
[
|
|
0.5555555555555556,
|
|
"#d8576b"
|
|
],
|
|
[
|
|
0.6666666666666666,
|
|
"#ed7953"
|
|
],
|
|
[
|
|
0.7777777777777778,
|
|
"#fb9f3a"
|
|
],
|
|
[
|
|
0.8888888888888888,
|
|
"#fdca26"
|
|
],
|
|
[
|
|
1.0,
|
|
"#f0f921"
|
|
]
|
|
],
|
|
"type": "surface"
|
|
}
|
|
],
|
|
"table": [
|
|
{
|
|
"cells": {
|
|
"fill": {
|
|
"color": "#EBF0F8"
|
|
},
|
|
"line": {
|
|
"color": "white"
|
|
}
|
|
},
|
|
"header": {
|
|
"fill": {
|
|
"color": "#C8D4E3"
|
|
},
|
|
"line": {
|
|
"color": "white"
|
|
}
|
|
},
|
|
"type": "table"
|
|
}
|
|
]
|
|
},
|
|
"layout": {
|
|
"annotationdefaults": {
|
|
"arrowcolor": "#2a3f5f",
|
|
"arrowhead": 0,
|
|
"arrowwidth": 1
|
|
},
|
|
"autotypenumbers": "strict",
|
|
"coloraxis": {
|
|
"colorbar": {
|
|
"outlinewidth": 0,
|
|
"ticks": ""
|
|
}
|
|
},
|
|
"colorscale": {
|
|
"diverging": [
|
|
[
|
|
0,
|
|
"#8e0152"
|
|
],
|
|
[
|
|
0.1,
|
|
"#c51b7d"
|
|
],
|
|
[
|
|
0.2,
|
|
"#de77ae"
|
|
],
|
|
[
|
|
0.3,
|
|
"#f1b6da"
|
|
],
|
|
[
|
|
0.4,
|
|
"#fde0ef"
|
|
],
|
|
[
|
|
0.5,
|
|
"#f7f7f7"
|
|
],
|
|
[
|
|
0.6,
|
|
"#e6f5d0"
|
|
],
|
|
[
|
|
0.7,
|
|
"#b8e186"
|
|
],
|
|
[
|
|
0.8,
|
|
"#7fbc41"
|
|
],
|
|
[
|
|
0.9,
|
|
"#4d9221"
|
|
],
|
|
[
|
|
1,
|
|
"#276419"
|
|
]
|
|
],
|
|
"sequential": [
|
|
[
|
|
0.0,
|
|
"#0d0887"
|
|
],
|
|
[
|
|
0.1111111111111111,
|
|
"#46039f"
|
|
],
|
|
[
|
|
0.2222222222222222,
|
|
"#7201a8"
|
|
],
|
|
[
|
|
0.3333333333333333,
|
|
"#9c179e"
|
|
],
|
|
[
|
|
0.4444444444444444,
|
|
"#bd3786"
|
|
],
|
|
[
|
|
0.5555555555555556,
|
|
"#d8576b"
|
|
],
|
|
[
|
|
0.6666666666666666,
|
|
"#ed7953"
|
|
],
|
|
[
|
|
0.7777777777777778,
|
|
"#fb9f3a"
|
|
],
|
|
[
|
|
0.8888888888888888,
|
|
"#fdca26"
|
|
],
|
|
[
|
|
1.0,
|
|
"#f0f921"
|
|
]
|
|
],
|
|
"sequentialminus": [
|
|
[
|
|
0.0,
|
|
"#0d0887"
|
|
],
|
|
[
|
|
0.1111111111111111,
|
|
"#46039f"
|
|
],
|
|
[
|
|
0.2222222222222222,
|
|
"#7201a8"
|
|
],
|
|
[
|
|
0.3333333333333333,
|
|
"#9c179e"
|
|
],
|
|
[
|
|
0.4444444444444444,
|
|
"#bd3786"
|
|
],
|
|
[
|
|
0.5555555555555556,
|
|
"#d8576b"
|
|
],
|
|
[
|
|
0.6666666666666666,
|
|
"#ed7953"
|
|
],
|
|
[
|
|
0.7777777777777778,
|
|
"#fb9f3a"
|
|
],
|
|
[
|
|
0.8888888888888888,
|
|
"#fdca26"
|
|
],
|
|
[
|
|
1.0,
|
|
"#f0f921"
|
|
]
|
|
]
|
|
},
|
|
"colorway": [
|
|
"#636efa",
|
|
"#EF553B",
|
|
"#00cc96",
|
|
"#ab63fa",
|
|
"#FFA15A",
|
|
"#19d3f3",
|
|
"#FF6692",
|
|
"#B6E880",
|
|
"#FF97FF",
|
|
"#FECB52"
|
|
],
|
|
"font": {
|
|
"color": "#2a3f5f"
|
|
},
|
|
"geo": {
|
|
"bgcolor": "white",
|
|
"lakecolor": "white",
|
|
"landcolor": "#E5ECF6",
|
|
"showlakes": true,
|
|
"showland": true,
|
|
"subunitcolor": "white"
|
|
},
|
|
"hoverlabel": {
|
|
"align": "left"
|
|
},
|
|
"hovermode": "closest",
|
|
"mapbox": {
|
|
"style": "light"
|
|
},
|
|
"paper_bgcolor": "white",
|
|
"plot_bgcolor": "#E5ECF6",
|
|
"polar": {
|
|
"angularaxis": {
|
|
"gridcolor": "white",
|
|
"linecolor": "white",
|
|
"ticks": ""
|
|
},
|
|
"bgcolor": "#E5ECF6",
|
|
"radialaxis": {
|
|
"gridcolor": "white",
|
|
"linecolor": "white",
|
|
"ticks": ""
|
|
}
|
|
},
|
|
"scene": {
|
|
"xaxis": {
|
|
"backgroundcolor": "#E5ECF6",
|
|
"gridcolor": "white",
|
|
"gridwidth": 2,
|
|
"linecolor": "white",
|
|
"showbackground": true,
|
|
"ticks": "",
|
|
"zerolinecolor": "white"
|
|
},
|
|
"yaxis": {
|
|
"backgroundcolor": "#E5ECF6",
|
|
"gridcolor": "white",
|
|
"gridwidth": 2,
|
|
"linecolor": "white",
|
|
"showbackground": true,
|
|
"ticks": "",
|
|
"zerolinecolor": "white"
|
|
},
|
|
"zaxis": {
|
|
"backgroundcolor": "#E5ECF6",
|
|
"gridcolor": "white",
|
|
"gridwidth": 2,
|
|
"linecolor": "white",
|
|
"showbackground": true,
|
|
"ticks": "",
|
|
"zerolinecolor": "white"
|
|
}
|
|
},
|
|
"shapedefaults": {
|
|
"line": {
|
|
"color": "#2a3f5f"
|
|
}
|
|
},
|
|
"ternary": {
|
|
"aaxis": {
|
|
"gridcolor": "white",
|
|
"linecolor": "white",
|
|
"ticks": ""
|
|
},
|
|
"baxis": {
|
|
"gridcolor": "white",
|
|
"linecolor": "white",
|
|
"ticks": ""
|
|
},
|
|
"bgcolor": "#E5ECF6",
|
|
"caxis": {
|
|
"gridcolor": "white",
|
|
"linecolor": "white",
|
|
"ticks": ""
|
|
}
|
|
},
|
|
"title": {
|
|
"x": 0.05
|
|
},
|
|
"xaxis": {
|
|
"automargin": true,
|
|
"gridcolor": "white",
|
|
"linecolor": "white",
|
|
"ticks": "",
|
|
"title": {
|
|
"standoff": 15
|
|
},
|
|
"zerolinecolor": "white",
|
|
"zerolinewidth": 2
|
|
},
|
|
"yaxis": {
|
|
"automargin": true,
|
|
"gridcolor": "white",
|
|
"linecolor": "white",
|
|
"ticks": "",
|
|
"title": {
|
|
"standoff": 15
|
|
},
|
|
"zerolinecolor": "white",
|
|
"zerolinewidth": 2
|
|
}
|
|
}
|
|
},
|
|
"title": {
|
|
"text": "Feature Correlation (Spearman, 154 pairs > 0.7)"
|
|
},
|
|
"width": 800
|
|
}
|
|
},
|
|
"image/png": "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"
|
|
},
|
|
"metadata": {},
|
|
"output_type": "display_data"
|
|
},
|
|
{
|
|
"name": "stdout",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"\n",
|
|
"High-correlation pairs (|rho| > 0.7): 154\n",
|
|
" cci_21 x bb_pctb_20: 0.987\n",
|
|
" vol_21d x natr_14: 0.962\n",
|
|
" rsi_7 x bb_pctb_20: 0.959\n",
|
|
" ret_252d x skip_recent_12_1: 0.956\n",
|
|
" vol_63d x vol_126d: 0.950\n",
|
|
" vol_63d x natr_14: 0.944\n",
|
|
" rsi_7 x stoch_k: 0.944\n",
|
|
" cci_14 x bb_pctb_20: 0.944\n",
|
|
" vol_126d x vol_252d: 0.943\n",
|
|
" cci_14 x cci_21: 0.939\n"
|
|
]
|
|
}
|
|
],
|
|
"source": [
|
|
"sample_df = features.filter(pl.col(\"timestamp\").is_in(dates[::5]))\n",
|
|
"# Rank in polars (fast), then Pearson on ranks = Spearman\n",
|
|
"corr_matrix = (\n",
|
|
" sample_df.select(feature_cols).select(pl.all().rank(method=\"average\")).to_pandas().corr()\n",
|
|
")\n",
|
|
"\n",
|
|
"high_corr_pairs = [\n",
|
|
" (f1, f2, corr_matrix.loc[f1, f2])\n",
|
|
" for f1, f2 in itertools.combinations(feature_cols, 2)\n",
|
|
" if abs(corr_matrix.loc[f1, f2]) > 0.7\n",
|
|
"]\n",
|
|
"\n",
|
|
"fig_corr = go.Figure(\n",
|
|
" go.Heatmap(\n",
|
|
" z=corr_matrix.values,\n",
|
|
" x=feature_cols,\n",
|
|
" y=feature_cols,\n",
|
|
" colorscale=\"RdBu_r\",\n",
|
|
" zmid=0,\n",
|
|
" zmin=-1,\n",
|
|
" zmax=1,\n",
|
|
" )\n",
|
|
")\n",
|
|
"fig_corr.update_layout(\n",
|
|
" title=f\"Feature Correlation (Spearman, {len(high_corr_pairs)} pairs > 0.7)\",\n",
|
|
" height=800,\n",
|
|
" width=800,\n",
|
|
")\n",
|
|
"fig_corr.show()\n",
|
|
"\n",
|
|
"print(f\"\\nHigh-correlation pairs (|rho| > 0.7): {len(high_corr_pairs)}\")\n",
|
|
"for f1, f2, rho in sorted(high_corr_pairs, key=lambda x: abs(x[2]), reverse=True)[:10]:\n",
|
|
" print(f\" {f1} x {f2}: {rho:.3f}\")"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "1dbdc7e0",
|
|
"metadata": {
|
|
"papermill": {
|
|
"duration": 0.004169,
|
|
"end_time": "2026-03-24T02:23:49.947748+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:23:49.943579+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"source": [
|
|
"**Expected vs surprising correlations**: Pairs like `ret_126d`/`ret_252d`\n",
|
|
"and `sharpe_126d`/`sharpe_252d` are correlated by construction (overlapping\n",
|
|
"lookback windows). Similarly, `vol_252d`/`vol_126d` share data. These\n",
|
|
"structural correlations are expected and handled by regularization (Ridge/Lasso)\n",
|
|
"in Ch11. Surprising high correlations (e.g., between volatility and momentum\n",
|
|
"features) would warrant investigation but are not observed here."
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "c30675f7",
|
|
"metadata": {
|
|
"papermill": {
|
|
"duration": 0.019532,
|
|
"end_time": "2026-03-24T02:23:49.971401+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:23:49.951869+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"source": [
|
|
"### Feature Family IC"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 23,
|
|
"id": "9cf697cc",
|
|
"metadata": {
|
|
"execution": {
|
|
"iopub.execute_input": "2026-04-09T01:07:59.649063Z",
|
|
"iopub.status.busy": "2026-04-09T01:07:59.648926Z",
|
|
"iopub.status.idle": "2026-04-09T01:07:59.655676Z",
|
|
"shell.execute_reply": "2026-04-09T01:07:59.655391Z"
|
|
},
|
|
"papermill": {
|
|
"duration": 0.010991,
|
|
"end_time": "2026-03-24T02:23:49.986599+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:23:49.975608+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"outputs": [
|
|
{
|
|
"name": "stdout",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"\n",
|
|
"Feature Family IC:\n",
|
|
"shape: (10, 5)\n",
|
|
"┌───────────────────┬───────────┬────────────┬────────────┬───────────────┐\n",
|
|
"│ family ┆ avg_ic ┆ avg_abs_ic ┆ n_features ┆ n_significant │\n",
|
|
"│ --- ┆ --- ┆ --- ┆ --- ┆ --- │\n",
|
|
"│ str ┆ f64 ┆ f64 ┆ u32 ┆ u32 │\n",
|
|
"╞═══════════════════╪═══════════╪════════════╪════════════╪═══════════════╡\n",
|
|
"│ distance ┆ 0.035135 ┆ 0.041219 ┆ 2 ┆ 1 │\n",
|
|
"│ volatility ┆ 0.033627 ┆ 0.036805 ┆ 10 ┆ 7 │\n",
|
|
"│ consistency ┆ 0.029384 ┆ 0.029384 ┆ 1 ┆ 1 │\n",
|
|
"│ momentum ┆ 0.007958 ┆ 0.023319 ┆ 14 ┆ 4 │\n",
|
|
"│ risk ┆ -0.018177 ┆ 0.018177 ┆ 2 ┆ 0 │\n",
|
|
"│ volume ┆ -0.014516 ┆ 0.014516 ┆ 1 ┆ 0 │\n",
|
|
"│ trend ┆ 0.005504 ┆ 0.012862 ┆ 4 ┆ 0 │\n",
|
|
"│ oscillator ┆ -0.007866 ┆ 0.008825 ┆ 8 ┆ 0 │\n",
|
|
"│ risk_adj_momentum ┆ 0.001112 ┆ 0.008073 ┆ 9 ┆ 0 │\n",
|
|
"│ regime ┆ 0.003418 ┆ 0.003418 ┆ 2 ┆ 0 │\n",
|
|
"└───────────────────┴───────────┴────────────┴────────────┴───────────────┘\n"
|
|
]
|
|
}
|
|
],
|
|
"source": [
|
|
"FAMILY_MAP = {\n",
|
|
" \"ret_\": \"momentum\",\n",
|
|
" \"skip_\": \"momentum\",\n",
|
|
" \"mom_accel\": \"momentum\",\n",
|
|
" \"sharpe_\": \"risk_adj_momentum\",\n",
|
|
" \"vol_\": \"volatility\",\n",
|
|
" \"natr_\": \"volatility\",\n",
|
|
" \"vol_ratio\": \"volume\",\n",
|
|
" \"obv_\": \"volume\",\n",
|
|
" \"rsi_\": \"oscillator\",\n",
|
|
" \"macd\": \"oscillator\",\n",
|
|
" \"adx_\": \"oscillator\",\n",
|
|
" \"cci_\": \"oscillator\",\n",
|
|
" \"stoch_\": \"oscillator\",\n",
|
|
" \"aroon\": \"oscillator\",\n",
|
|
" \"sma_\": \"trend\",\n",
|
|
" \"ema_\": \"trend\",\n",
|
|
" \"bb_\": \"trend\",\n",
|
|
" \"chop_\": \"regime\",\n",
|
|
" \"hurst_\": \"regime\",\n",
|
|
" \"regime\": \"regime\",\n",
|
|
" \"yield_\": \"regime\",\n",
|
|
" \"corr_spy\": \"regime\",\n",
|
|
" \"max_dd\": \"risk\",\n",
|
|
" \"pct_positive\": \"consistency\",\n",
|
|
" \"dist_\": \"distance\",\n",
|
|
"}\n",
|
|
"\n",
|
|
"\n",
|
|
"family_ic = (\n",
|
|
" eval_summary.with_columns(\n",
|
|
" pl.coalesce(\n",
|
|
" [\n",
|
|
" pl.when(pl.col(\"feature\").str.starts_with(prefix))\n",
|
|
" .then(pl.lit(family))\n",
|
|
" .otherwise(pl.lit(None))\n",
|
|
" for prefix, family in FAMILY_MAP.items()\n",
|
|
" ]\n",
|
|
" + [\n",
|
|
" pl.when(pl.col(\"feature\").str.ends_with(\"_rank\"))\n",
|
|
" .then(pl.lit(\"rank\"))\n",
|
|
" .otherwise(pl.lit(\"other\"))\n",
|
|
" ]\n",
|
|
" ).alias(\"family\")\n",
|
|
" )\n",
|
|
" .group_by(\"family\")\n",
|
|
" .agg(\n",
|
|
" pl.col(\"ic_mean\").mean().alias(\"avg_ic\"),\n",
|
|
" pl.col(\"ic_mean\").abs().mean().alias(\"avg_abs_ic\"),\n",
|
|
" pl.len().alias(\"n_features\"),\n",
|
|
" pl.col(\"significant_fdr05\").sum().alias(\"n_significant\"),\n",
|
|
" )\n",
|
|
" .sort(\"avg_abs_ic\", descending=True)\n",
|
|
")\n",
|
|
"print(\"\\nFeature Family IC:\")\n",
|
|
"print(family_ic)"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "3eeb2765",
|
|
"metadata": {
|
|
"papermill": {
|
|
"duration": 0.004139,
|
|
"end_time": "2026-03-24T02:23:49.994954+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:23:49.990815+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"source": [
|
|
"**Evaluation Results**:\n",
|
|
"\n",
|
|
"- HAC adjustment deflates t-statistics relative to naive, reflecting the\n",
|
|
" autocorrelation in overlapping 21-day returns\n",
|
|
"- BH-FDR correction further culls features that appeared significant only\n",
|
|
" due to multiple testing\n",
|
|
"- Feature correlation heatmap reveals redundancy clusters that downstream\n",
|
|
" models should address through regularization or explicit selection"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "d2e75068",
|
|
"metadata": {
|
|
"papermill": {
|
|
"duration": 0.004111,
|
|
"end_time": "2026-03-24T02:23:50.003246+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:23:49.999135+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"source": [
|
|
"## 7. Results Collection"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 24,
|
|
"id": "73b6cfae",
|
|
"metadata": {
|
|
"execution": {
|
|
"iopub.execute_input": "2026-04-09T01:07:59.656652Z",
|
|
"iopub.status.busy": "2026-04-09T01:07:59.656576Z",
|
|
"iopub.status.idle": "2026-04-09T01:07:59.658716Z",
|
|
"shell.execute_reply": "2026-04-09T01:07:59.658461Z"
|
|
},
|
|
"papermill": {
|
|
"duration": 0.007652,
|
|
"end_time": "2026-03-24T02:23:50.015111+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:23:50.007459+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"outputs": [],
|
|
"source": [
|
|
"top_features = (\n",
|
|
" eval_summary.filter(pl.col(\"significant_fdr05\"))\n",
|
|
" .sort(pl.col(\"ic_mean\").abs(), descending=True)\n",
|
|
" .head(5)[\"feature\"]\n",
|
|
" .to_list()\n",
|
|
")\n",
|
|
"family_ic_dict = {row[\"family\"]: round(row[\"avg_ic\"], 4) for row in family_ic.to_dicts()}"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "e6fb58e1",
|
|
"metadata": {
|
|
"papermill": {
|
|
"duration": 0.004085,
|
|
"end_time": "2026-03-24T02:23:50.023353+00:00",
|
|
"exception": false,
|
|
"start_time": "2026-03-24T02:23:50.019268+00:00",
|
|
"status": "completed"
|
|
},
|
|
"tags": []
|
|
},
|
|
"source": [
|
|
"## Key Takeaways\n",
|
|
"\n",
|
|
"1. **Skip-recent momentum** (12-1 and 6-1) is the most important literature-standard\n",
|
|
" feature that was missing -- it avoids short-term reversal contamination\n",
|
|
"2. **Cross-sectional ranks** limited to 3 primary features (ret_126d, sharpe_126d,\n",
|
|
" vol_63d) to reduce multicollinearity\n",
|
|
"3. **MA ratio proliferation** reduced from 8 to 3 (SMA 50/200, EMA 26)\n",
|
|
"4. **Point-in-time eligibility** enforced on feature matrix -- no phantom ETFs\n",
|
|
" inflating early-period cross-section breadth\n",
|
|
"5. **SPY-TLT 63d correlation** added as cross-asset regime indicator\n",
|
|
"6. **Yield curve z-score** provides continuous regime signal (better than binary)\n",
|
|
"7. **IC evaluation with HAC + FDR** validates feature significance -- overlapping\n",
|
|
" labels inflate naive t-stats; multiple testing further culls spurious features\n",
|
|
"8. **Gap: regime dynamics**: Cross-sectional features capture relative\n",
|
|
" positioning but not market-level regime shifts. Ch9 adds data-driven\n",
|
|
" regime detection (HMM) and memory-preserving transforms (FFD) that\n",
|
|
" complement the rule-based regime indicators here.\n",
|
|
"\n",
|
|
"**Next**: `04_temporal.py` fits HMM regimes and fractional differencing models."
|
|
]
|
|
}
|
|
],
|
|
"metadata": {
|
|
"jupytext": {
|
|
"cell_metadata_filter": "-all"
|
|
},
|
|
"kernelspec": {
|
|
"display_name": "Python 3 (ipykernel)",
|
|
"language": "python",
|
|
"name": "python3"
|
|
},
|
|
"language_info": {
|
|
"codemirror_mode": {
|
|
"name": "ipython",
|
|
"version": 3
|
|
},
|
|
"file_extension": ".py",
|
|
"mimetype": "text/x-python",
|
|
"name": "python",
|
|
"nbconvert_exporter": "python",
|
|
"pygments_lexer": "ipython3",
|
|
"version": "3.14.3"
|
|
},
|
|
"papermill": {
|
|
"default_parameters": {},
|
|
"duration": 96.852286,
|
|
"end_time": "2026-03-24T02:23:51.242757+00:00",
|
|
"environment_variables": {},
|
|
"exception": null,
|
|
"input_path": "case_studies/etfs/03_financial_features.ipynb",
|
|
"output_path": "/tmp/ml4t_3987651_03_financial_features.ipynb",
|
|
"parameters": {
|
|
"GARCH_MIN_OBS": 50,
|
|
"HMM_N_RESTARTS": 1,
|
|
"MAX_SAMPLE_ASSETS": 3,
|
|
"MAX_SYMBOLS": 5,
|
|
"MIN_OBS": 10,
|
|
"MIN_TRAIN_BARS": 10,
|
|
"N_RESTARTS": 1,
|
|
"START_DATE": "2024-06-01"
|
|
},
|
|
"start_time": "2026-03-24T02:22:14.390471+00:00",
|
|
"version": "2.7.0"
|
|
}
|
|
},
|
|
"nbformat": 4,
|
|
"nbformat_minor": 5
|
|
}
|