Fuzzy C-means Clustering Based on the Multivariable Weighted Attributes

Considering the difference of feature attributes, a new clustering algorithm based on the importance of attributes is proposed according to multivariable fuzzy c-means clustering method.In order to verify the validity of new method, the proposed algorithm is analyzed and compared with principal component comprehensive analysis and attributes weighted by rough set theory on six UCI datasets.The experimental results showed that this method has wider scope of application and stability.Moreover, the clustering efficiency increase when datasets have further average distance among cluster centers.