计算机应用与软件2012，Vol.29Issue(9)：146-147,174,3.DOI:10.3969/j.issn.1000-386x.2012.09.038
属性相似度在聚类算法中的有效性研究
ON EFFECTIVENESS OF ATTRIBUTE SIMILARITY IN CLUSTERING ALGORITHM
摘要
Abstract
Euclidean distance identifies the distinctions of different property (i.e. each index or variables) of individuals but neglects the importance of them. To solve this, considering the geometric structure characteristics and the individual property of data and in combination with Mahalanobis distance, we present a new property similarity metric method and the new clustering validity function, and improve the hierarchical clustering algorithm using Euclidean distance. Experiment results show that the improved clustering algorithm can enhance the speed and quality of clustering, and is an effective clustering method.关键词
相似性/聚类算法/有效性/度量方法Key words
Similarity/Clustering algorithm/Validity/Metric method分类
信息技术与安全科学引用本文复制引用
刘明术,方宏彬,张建,孙启林..属性相似度在聚类算法中的有效性研究[J].计算机应用与软件,2012,29(9):146-147,174,3.基金项目
安徽省教育厅自然科学基金项目(05010428) （05010428）
安徽大学人才队伍建设项目. （）
