Attribute Weighted Optimization of Fuzzy C-Means Clustering Algorithm.

According to the standard fuzzy C-means clustering algorithm performed poor in the clustering effect during the clustering process. This paper presents an objective function optimization based on the attribute weighted and the objective function optimization. Firstly, use a little prior knowledge as the labeled sample. These calibrated samples information are used as the prior knowledge, and then use the self-creating method and the statistical characteristics based on the data to optimize the attribute weights in FCM algorithm, and then introduce the kernel function to improve the search ability of the fuzzy C-means clustering algorithm, simplify both the clustering center and membership matrix with lagrange multiplier approach. The simulation experiment shows that, contrast to the original algorithm and K-means clustering algorithm, the fuzzy C-means clustering algorithm optimized by the attribute weighted method presented in this paper has a better clustering effect. [ABSTRACT FROM AUTHOR]