SemiYieldNet Results Dataset: Benchmark Predictions, Uncertainty Estimates, Calibration Metrics, Statistical Tests, and SHAP Explanations for Semiconductor Yield Prediction

This dataset contains the experimental results generated by SemiYieldNet, an uncertainty-aware stacked ensemble framework for semiconductor yield prediction using the SECOM benchmark dataset. The repository includes model prediction outputs, performance metrics, ablation study results, statistical significance tests, Monte Carlo Dropout uncertainty estimates (epistemic variance and predictive entropy), SHAP feature importance rankings, and runtime statistics. The dataset enables full reproduction of the evaluation, calibration, uncertainty quantification, and interpretability analyses reported in the associated manuscript.