Simple hybrid and incremental postpruning techniques for rule induction

This document presents a new hybrid pruning technique and an incremental postpruning technique for rule induction algorithms. The techniques aim to reduce model complexity, improve predictive accuracy, and minimize overfitting. An empirical evaluation found that the techniques enabled RULES-7, a rule induction algorithm, to outperform other classification techniques while also producing more accurate and faster classifiers.