关键词: 中心权重, 模糊划分, 关系数据, 非相似性

Abstract: It is apparently inadequate that k center algorithm uses only one point to represent the whole class, which will definitely affect the accuracy of clustering results. Therefore, a weighted fuzzy clustering algorithm for relational data with multiple medoids was proposed. In the proposed algorithm, multiple objects in each cluster carried different weights called medoids weights to represent their degrees of representativeness in that cluster. This algorithm can make each cluster to be represented by multiple objects instead of only one object. The experimental results show that the proposed algorithm can capture the underlying structures of the data more accurately and provide richer information for the description of the resulting clusters.

Key words: weighted medoid, fuzzy partition, relational data, dissimilarity