Extraction of Rules from Coronary Heart Disease Database Using Automatically Defined Groups

Abstract

Much of the research on extracting rules from a large amount of data has focused on the extraction of a general rule that satisfies as many data as possible. In the field of health care where people’s lives are at stake, the exceptional rules for rare cases are also important. In this paper, we describe the knowledge acquisition from data containing such multiple rules. We consider that a multi-agent approach is effective for extracting multiple rules. In order to realize this approach, we propose a new method using an improved Genetic Programming method, Automatically Defined Groups (ADG). By using this method, the clustering of data is performed by sharing roles among agents, and each agent takes charge of rule extraction in the assigned data. As a result, all rules are extracted by multi-agent cooperation. We applied this method to coronary heart disease databases, and showed the effectiveness of this method.