An Improvement Clustering Algorithms Based on Max-t1 Norm
MAX—t1范数聚类算法的改进

A comparative study between transitive closure fuzzy clustering algorithm and the fuzzy clustering algorithm based on t_1-norm with fuzzy matrix are presented. The study shows that for describing the fuzzy relation of two elements max-t_1 fuzzy transitive relation is more critical than that of max-min fuzzy transitive relation. That is the reason why t_1-norm dose not has max-t transitivity. Besides, in this paper a new algorithm to determine equivalent classes is given. With the new algorithm the fuzzy clustering algorithm based on t_1-norm with fuzzy matrix is more reasonable. Experiments show the effectiveness of the new algorithm.