Variation approaches to feature-weight selection and application to fuzzy clustering

In statistics field, variation plays an important role. This is because greater variations in some features of data can provide more important information. Therefore, in this paper, we use this idea to select feature-weights in data. The proposed approach is simple to compute and interpret for feature-weights selection. Compared with the feature-weights proposed by Wang et al., Modha and Spangler, Pal et al. & Basak et al., we find that the proposed method provides a better clustering performance for the Iris data and color image segmentation and also has lower computational complexity..