Supplementary Data for "Text Sentiment Classification Models: Predictive Performance, Calibration, and LIME Explanations"

This record contains five CSV files with supplementary data for the article “Text Sentiment Classification Models: Predictive Performance, Calibration, and LIME Explanations.” Supplementary Table S1 reports model checkpoints, fixed hyperparameters, validation settings, computing environment, and fit times. Supplementary Table S2 provides test-set performance metrics for 21 classifiers. Supplementary Table S3 contains bootstrap confidence intervals for Macro-F1, Brier score, and ECE. Supplementary Table S4 reports all 210 Holm-adjusted pairwise McNemar test results. Supplementary Table S5 presents token weights and record counts for the selected LIME explanations. The files are provided in CSV format.