Feature Description Systems for Clusters by Using Logical Rule Generations Based on the Genetic Programming and Its Applications to Data Mining

摘要

This paper deals with the realization of retrieval and feature description systems for clusters by using logical rule generations based on the Genetic Programming (GP). At first, whole data is divided into several clusters and the rules are improved based the GP. The fitness of individuals is defined in proportion to the hits of corresponding logical expression to the samples in targeted cluster c, but also in inversely proportion to the hits outside the cluster c. The GP method is applied to various real world data by showing effective performance compared to conventional methods.