Research of Algorithm Based on Improved Self-Organization Neural Network of Fuzzy Clustering

This paper introduced an improved fuzzy SOFM clustering algorithm. In the initialization phase, by the way of subtractive clustering, optimized initial weights of network and determined the number of clusters. To verify the effectiveness of the algorithm, this algorithm will be applied to web log mining. Experimental results showed that the improved fuzzy SOFM neural network training speed and convergence results have improved to some extent, and for a variety of users interested in mining provides a feasible approach.