A rule induction algorithm for continuous data using analysis of variance

Knowledge acquisition continues to be a challenging and time consuming task in building decision support systems. Among the dominant methods, rule induction algorithms such as ID3 and C4.5 are widely used to extract rules from examples. The thrust of these algorithms is how they discriminate the given attributes based on information measure for building and determining the nodes in the decision tree. In particular, the main focus of these algorithms is on how to select the most appropriate attribute at each level of the decision tree process. This paper proposes an algorithm for rule induction for continuous data. The proposed algorithm uses an analysis of variance criterion for information measure in discriminating the given attributes for building the decision tree for continuous data.