README.md

metadata

license: other
tags:
  - missing-data
  - imputation
  - benchmark
  - healthcare

Missing-Data Prediction Benchmark — results

Two experiments on medical datasets under simulated missingness (MCAR/MAR/MNAR, rates 0.1–0.5, seeds 42/1/7), reported as mean [95% CI].

This release contains the complete Information-Acquisition benchmark at a uniform missing rate of 0.3. Scenario 2 includes local AFA-engine results, curves, seed-aggregated CI tables, qualitative examples, classification AUROC, and per-engine cached/uncached token accounting across five classification datasets (arrhythmia, breast_cancer, gallstone, heart_disease, parkinsons) and four regression datasets (diabetes, nhanes, who_life_expectancy, medical_cost).

Scenario 1 uses gpt-5.6-terra to select features and predict directly. It is complete for the same classification and regression dataset sets. Classification Scenario 1 reports accuracy, F1, precision, and recall are preserved from the original label-only query. Classification AUROC comes from a separate probability replay with the original selection path unchanged; this mixed provenance is noted in the A/B report. Regression Scenario 1 predicts a continuous numeric target and reports RMSE, MAE, and R2. MIMIC Scenario 1 is absent because credentialed MIMIC-derived records are not sent to third-party APIs.

No patient-level records, raw MIMIC data, trained models, or LLM traces containing MIMIC features are published.

Classification Scenario 1 vs Scenario 2

dataset	mech	LLM acc	AFA acc	LLM F1	AFA F1	LLM AUROC	AFA AUROC	best AFA	LLM tokens	S2 cached	S2 uncached	↓ cached	↓ uncached
arrhythmia	MAR	0.458	0.736	0.44	0.734	0.494	0.772	random_forest	10,516,156	31,992	417,730	99.7%	96.0%
arrhythmia	MCAR	0.473	0.74	0.456	0.738	0.431	0.784	random_forest	10,544,703	31,992	404,444	99.7%	96.2%
arrhythmia	MNAR	0.513	0.707	0.495	0.704	0.423	0.771	random_forest	10,926,848	31,992	414,435	99.7%	96.2%
breast_cancer	MAR	0.921	0.977	0.921	0.977	0.979	0.996	smart	348,937	4,835	56,524	98.6%	83.8%
breast_cancer	MCAR	0.927	0.98	0.927	0.979	0.979	0.996	mim_smim	373,315	4,191	13,165	98.9%	96.5%
breast_cancer	MNAR	0.927	0.974	0.927	0.974	0.983	0.995	smart	348,073	4,835	67,905	98.6%	80.5%
gallstone	MAR	0.495	0.771	0.483	0.769	0.543	0.838	hist_gb	823,103	5,990	50,355	99.3%	93.9%
gallstone	MCAR	0.495	0.755	0.482	0.755	0.484	0.814	mim_smim	830,642	6,043	43,630	99.3%	94.7%
gallstone	MNAR	0.51	0.771	0.499	0.77	0.49	0.817	random_forest	814,237	6,043	48,555	99.3%	94.0%
heart_disease	MAR	0.7	0.844	0.698	0.844	0.78	0.903	smart	197,635	2,068	18,610	99.0%	90.6%
heart_disease	MCAR	0.694	0.856	0.693	0.855	0.672	0.893	mim_smim	208,505	2,068	11,775	99.0%	94.4%
heart_disease	MNAR	0.694	0.833	0.693	0.833	0.768	0.899	smart	198,665	2,068	18,187	99.0%	90.8%
parkinsons	MAR	0.761	0.897	0.776	0.889	0.841	0.961	hist_gb	159,996	3,368	19,475	97.9%	87.8%
parkinsons	MCAR	0.752	0.889	0.767	0.887	0.87	0.931	hist_gb	155,967	3,322	15,660	97.9%	90.0%
parkinsons	MNAR	0.744	0.872	0.759	0.873	0.836	0.948	xgboost	164,584	3,373	14,908	98.0%	90.9%

How to read (no download needed)

Every CSV opens in HuggingFace's Dataset Viewer (sortable, in-browser):

• experiment1_prediction/prediction_classification_long.csv / prediction_regression_long.csv — one tidy row per (method, dataset, mechanism, rate, seed) with all metrics + status. This is the clean, flat view of every run (no folder-diving).
• experiment1_prediction/classification_mean_ci.csv / regression_mean_ci.csv — the same, collapsed to mean [95% CI] over seeds (blank cells carry a Status).
• experiment2_acquisition/acquisition_ab_report.csv — matched-sample classification Scenario 1 vs Scenario 2 accuracy and token comparison.
• experiment2_acquisition/curves_long.csv — per-budget Scenario 2 trajectories and Scenario 1 trajectories for both tasks.
• experiment2_acquisition/acquisition_curves_ci.csv / acquisition_final_ci.csv — seed-aggregated metrics for both scenarios.
• experiment2_acquisition/scenario2_method_token_accounting.csv — per-engine cached and uncached Scenario-2 token accounting.
• DATA_DICTIONARY.md — column definitions and status meanings.
• raw_prediction/ — per-cell raw Exp-1 metric JSON (in-scope datasets, canonical layout, see raw_prediction/README_LATEST.md).
• raw_acquisition/ — per-cell raw complete Scenario-1 LLM and Scenario-2 AFA outputs (no model weights).

Data-use / MIMIC-III notice

One regression dataset (mimic_lengthofstay) is derived from MIMIC-III, credentialed under the PhysioNet Data Use Agreement. This repository contains only aggregate derived metrics — no patient-level records, no raw MIMIC data, and no models trained on MIMIC.

Generated 2026-09-09 by scripts/reporting/prepare_hf_upload.py.