Weighted Fuzzy C-Means Clustering Based on Double Coding Genetic Algorithm

We propose a double coding scheme in genetic algorithm (GA) and apply it to the fuzzy features-weighting clustering problems. Each individual consists of two segments of codes for cluster centers and feature weights. The two segments are evolved simultaneously in the clustering process. A modified clustering objective function is defined. A weighted fuzzy c-means operator and a feature weights learning operator are designed to guide computing cluster centers and feature weights in an individual respectively. On the basis of the above work, a novel weighed fuzzy c-means clustering algorithm based on double coding GA is advanced.