CTCHAID: Extending the application of the consolidation methodology

Abstract

The consolidation process, originally applied to the C4.5 tree induction algorithm, improved its discriminating capacity and stability. Consolidation creates multiple samples and builds a simple (nonmultiple) classifier by applying the ensemble process during the model construction times. A benefit of consolidation is that the understandability of the base classifier is kept. The work presented aims to show the consolidation process can improve algorithms other than C4.5 by applying the consolidation process to another algorithm, CHAID*. The consolidation of CHAID*, CTCHAID, required solving the handicap of consolidating the value groupings proposed by each CHAID* tree for discrete attributes. The experimentation is divided in three classification contexts for a total of 96 datasets. Results show that consolidated algorithms perform robustly, ranking competitively in all contexts, never falling into lower positions unlike most of the other 23 rule inducting algorithms considered in the study. When performing a global comparison consolidated algorithms rank first.