The Hybrid of Genetic Algorithms and K-Prototypes Clustering Approach for Classification
Chaochang Chiu, Huaichun Chi, Rueijiau Sung, Ju-Yun Yuang · 2010
This study proposes a novel classification technique of GA/k-prototypes in combination with a genetic algorithm to take the advantage of k-prototypes clustering mechanism for supporting the classification purpose. A genetic algorithm is used to adjust the weight applied to input attributes in order to enable a majority of the data records in each cluster to be with the same outcome class. We conduct three experiments with the GA/k-prototypes classification algorithm using UCI repository data sets. The experimental results show that the proposed approach can achieve superior classification performance than other commonly used data mining approaches.