FCM Clustering Algorithm Based on Laplacian Coefficient Optimized Objective Function

FCM (Fuzzy c-means) clustering algorithm can be used to build sample generic uncertainty description. In this paper, we propose a FCM clustering algorithm based on the Laplacian Coefficient optimization objective function. The Laplacian Coefficient is introduced in the objective function, the structure of information between the object into the weight, thus improve the quality and efficiency of algorithm, and then through the compactness and separability measure two parts to optimize clustering validity, and use the method of maximum effectiveness will function for standardization, in order to improve the antinoise performance of improved algorithm. In the UCI standard data sets of Iris and Wine simulation experiments show that the proposed improved FCM algorithm compared with standard algorithm has more accurate clustering effect, and less affected by noise and higher robustness.