Kernel fuzzy C-means algorithm based on distribution density and its application in fault diagnosis
Xinmin Tao · Zhendong yu chongji · 2009
To solve the problem of sensibility of traditional kernel fuzzy C-means algorithm (KFCM) to outliers and noises in training sets, a novel kernel fuzzy C-means algorithm based on distribution density around samples combined with maximum variance between clusters was proposed. In the proposed method, the value of distribution density around samples was used as weighted values according to the feature of sample distribution to overcome the shortcomings of sensibility to and outliers noises. The maximum variance between clusters was applied to segment the samples’ distribution density vector, and the segmentation results were used as the initial centers of the proposed KFCM algorithm, which overcomes the problem of sensibility to initial values. The proposed method can be applied to analyze the samples’ contribution to clustering performance. The experimental results with various real data sets illustrate the effectiveness of the algorithm. The proposed method can be also utilized in fault diagnosis field which outperforms traditional cluster methods.