Tag
This paper introduces a statistically grounded framework using Bernoulli Naïve Bayes and χ²-guided binarization for interpretable clinical classification, achieving competitive AUC scores on benchmark datasets while providing explicit decision rules and calibrated risk estimates.