An Exploration of a Set of Entropy-Based Hybrid Splitting Methods for Splitting Methods for

摘要

Decision tree (DT) induction is among the more popular of the data mining techniques. An important component of DT induction algorithms is the splitting method, with the most commonly used method being based on the Conditional Entropy family. However, it is well known that there is no single splitting method that will give the best performance for all problem instances. In this paper, we develop and explore hybrid splitting methods from two entropy-based families: the Conditional Entropy family and another family that is based on the Class-Attribute Mutual Information (CAMI). We compare conventional splitting methods based on single measures with hybrid splitting methods based on multiple measures. The results suggest that the hybrid methods could be competitive in terms of classification accuracy and are thus worthy of future research.