模糊聚类分析的传递方法

Transfer Algorithm to Fuzzy Clustering Analysis

针对常规模糊聚类分析在改造模糊相似关系时存在的复杂矩阵幂乘运算问题，提出了模糊聚类分析的传递算法，通过设置置信水平λ，直接从模糊相似关系获得聚类结果，并证明该方法与传递闭包法等价，但时间复杂度和空间复杂度都要远远小于传递闭包法．

This paper proposes a new transfer algorithm to fuzzy clustering analysis with respect to the problem of complex operation of matrix synthesis when reconstructing fuzzy analogical relation to fuzzy equivalence relation .By setting the threshold to λ, we get the clustering results from fuzzy analogical relation directly. The transfer algorithm is equivalent to transfer closure algorithm, but its asymptotic time complexity and space complexity is superior to the latter.