LOBench-A-share-processed / code /experiments /tbd_revision /scripts /summarize_label_threshold_sensitivity.py
| #!/usr/bin/env python3 | |
| """Summarize how the predefined A-share thresholds change natural labels.""" | |
| from __future__ import annotations | |
| import argparse | |
| import csv | |
| import statistics | |
| from collections import defaultdict | |
| from pathlib import Path | |
| def percentile(values: list[float], probability: float) -> float: | |
| ordered = sorted(values) | |
| position = probability * (len(ordered) - 1) | |
| lower = int(position) | |
| upper = min(lower + 1, len(ordered) - 1) | |
| weight = position - lower | |
| return ordered[lower] * (1 - weight) + ordered[upper] * weight | |
| def parse_args() -> argparse.Namespace: | |
| parser = argparse.ArgumentParser(description=__doc__) | |
| parser.add_argument("--input", type=Path, required=True) | |
| parser.add_argument("--output", type=Path, required=True) | |
| parser.add_argument("--split", default="test") | |
| return parser.parse_args() | |
| def main() -> None: | |
| args = parse_args() | |
| grouped: dict[float, list[dict[str, str]]] = defaultdict(list) | |
| with args.input.open(newline="", encoding="utf-8") as handle: | |
| for row in csv.DictReader(handle): | |
| if row["split"] == args.split: | |
| grouped[float(row["threshold"])].append(row) | |
| rows = [] | |
| for threshold, cases in sorted(grouped.items()): | |
| stationary_rates = [ | |
| int(row["stationary"]) / int(row["n_windows"]) for row in cases | |
| ] | |
| pooled_n = sum(int(row["n_windows"]) for row in cases) | |
| pooled_down = sum(int(row["down"]) for row in cases) | |
| pooled_stationary = sum(int(row["stationary"]) for row in cases) | |
| pooled_up = sum(int(row["up"]) for row in cases) | |
| rows.append( | |
| { | |
| "split": args.split, | |
| "threshold": threshold, | |
| "n_stocks": len(cases), | |
| "pooled_windows": pooled_n, | |
| "pooled_down_rate": pooled_down / pooled_n, | |
| "pooled_stationary_rate": pooled_stationary / pooled_n, | |
| "pooled_up_rate": pooled_up / pooled_n, | |
| "stock_stationary_rate_mean": statistics.fmean(stationary_rates), | |
| "stock_stationary_rate_q1": percentile(stationary_rates, 0.25), | |
| "stock_stationary_rate_median": statistics.median(stationary_rates), | |
| "stock_stationary_rate_q3": percentile(stationary_rates, 0.75), | |
| } | |
| ) | |
| args.output.parent.mkdir(parents=True, exist_ok=True) | |
| with args.output.open("w", newline="", encoding="utf-8") as handle: | |
| writer = csv.DictWriter(handle, fieldnames=list(rows[0])) | |
| writer.writeheader() | |
| writer.writerows(rows) | |
| markdown = args.output.with_suffix(".md") | |
| lines = [ | |
| "# Predefined label-threshold sensitivity", | |
| "", | |
| f"Natural `{args.split}` split; no resampling or model-selection outcome is used.", | |
| "", | |
| "| threshold | stocks | pooled windows | down | stationary | up | stock stationary median [IQR] |", | |
| "|---:|---:|---:|---:|---:|---:|---:|", | |
| ] | |
| for row in rows: | |
| lines.append( | |
| f"| {row['threshold']:.4f} | {row['n_stocks']} | " | |
| f"{row['pooled_windows']} | {row['pooled_down_rate']:.3f} | " | |
| f"{row['pooled_stationary_rate']:.3f} | {row['pooled_up_rate']:.3f} | " | |
| f"{row['stock_stationary_rate_median']:.3f} " | |
| f"[{row['stock_stationary_rate_q1']:.3f}, " | |
| f"{row['stock_stationary_rate_q3']:.3f}] |" | |
| ) | |
| markdown.write_text("\n".join(lines) + "\n", encoding="utf-8") | |
| print(f"wrote {args.output} and {markdown}") | |
| if __name__ == "__main__": | |
| main() | |