基于新相异度量的模糊K-Modes聚类算法

Fuzzy K-modes Clustering Algorithm Based on a New Dissimilarity Measure

提出了一种基于新相异度量的模糊K-Modes算法。该算法假定不同属性对聚类结果有不同程度的影响,定义了新的属性值函数,以基于划分相似度的聚类精确度作为聚类结果的评价准则。通过真实数据的实验结果表明,新的基于相异度量的模糊K-Modes算法比传统的模糊K-Modes算法有更好的聚类效果。

A fuzzy K-Modes clustering algorithm based on New Dissimilarity Measure is presented for the different contribution of each attribute of the data set to the clustering. With a new function of the attribute value, the clustering accuracy based on the partition similitude is used to evaluate the clustering result. Experimental results on real life data show that the new fuzzy k-modes algorithm is superior to the standard fuzzy k-modes algorithm with respect to clustering accuracy.