Dataset-Specific Bootstrap-Stability Weighting for Calibrated and Clinically Useful Ensemble Prediction in Medical Diagnosis
Bibliographic record
Abstract
Background: Ensemble machine-learning models often perform well within a single medical dataset yet lose discrimination, calibration, and decision usefulness under dataset shift. Objective: To develop and evaluate Bootstrap-Guided Optimization System (BOOTMED), a bootstrap-guided framework that learns dataset-specific weights from resampling stability to fuse probabilistic predictions, targeting discrimination, calibration, and decision-analytic utility simultaneously. Methods: Four heterogeneous UCI medical datasets were analyzed (Chronic Kidney Disease; CKD, diabetes, heart disease, breast cancer). Base learners were k-nearest neighbors, random forest (RF), Gaussian naïve Bayes, and complement naïve Bayes. BOOTMED estimated stability-derived weights over 500 bootstrap resamples and aggregated model probabilities. Performance was compared with equal-weight voting and stacking using balanced accuracy and ROC-AUC, calibration error (Brier/ECE), and decision curve analysis. Results: BOOTMED outperformed equal-weight voting and the best single model across all datasets, improving balanced accuracy by approximately 0.7-2.3 percentage points (adjusted p<0.05). Calibration error decreased (lower Brier/ECE), and decision curve analysis showed consistent positive net benefit across clinically relevant thresholds (0.10-0.50). Transferring weights between datasets reduced performance, supporting dataset-specific optimization. Conclusion: Bootstrap-guided, dataset-specific weighting can improve discrimination, calibration, and clinical net benefit across heterogeneous medical datasets, offering a simple and reproducible ensembling strategy for diagnostic prediction.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.017 | 0.048 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.000 | 0.002 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".