基于子集测度的模糊加权指数进化计算方法

Evolution computing method of fuzzy weighted exponent based on subset measuring

为解决在模糊C-均值聚类应用过程中,模糊加权指数的确定缺乏理论依据和有效评价方法的问题,提出了一种基于子集测度的模糊加权指数进化计算方法。根据子集测度理论定义聚类有效性函数,在聚类过程中通过循环进化迭代计算聚类的有效性,并将其值反馈到模糊加权指数的变化中,使收敛到一个稳定解。通过实验得到的加权指数符合预期的结果,理论分析和实验结果表明了该方法的有效性。

To deal with the problem of lacking theoretical foundation and effective methodology in the application process of Fuzzy C-Means clustering, a fuzzy weighted exponent evolution computing method is put forward based on subset measuring. Firstly, clustering valid fimction is defined based on subset measuring theory. Then in the clusering process the validity of clustering is calculated by evolution iteration loop and it is feed back to the change of fuzzy weighted exponent ＂m＇, then the fuzzy weighted exponent ＂m＂ will be converged to a stable optimum relation. The weighted exponent by experiments meets the expected results. The method is valid by theo-retical analysis and experiment.