Immune Algorithm for Supervised Clustering

摘要：

This paper centers on a navel data mining technique we term immune supervised clustering. Unlike traditional clustering, immune supervised clustering assumes that the examples are classified by immune algorithm. The goal of immune supervised clustering algorithm(ISCA) is to identify class-uniform clusters that have high probability densities. The experimental results suggest that ISCA, although runtime intensive, finds the best clusters in almost all experiments conducted.

关键词： Supervised clustering Immune algorithm Clustering for classification.

作者: Lifang Xu Hongwei Mo Kejun Wang

作者单位: Automation College, Harbin Engineering Univeristy

会议类型: 国际会议

会议名称: Firth IEEE International Conference on Cognitive Informatics(第五届认知信息国际会议)

会议地点: 北京

会议语种:英文

页码: 953-958

在线出版日期: 2006-07-17（万方平台首次上网日期，不代表论文的发表时间）