竞争式模糊聚类学习在函数逼近中的应用

Competitive Fuzzy Clustering with Application in Function Approach

本文通过对RCA算法中遗忘函数的修正,抑制了类间竞争迭代中的病态发散,从而实现了算法的稳健收敛。用该算法分析数据,采用TSK模糊模型对函数进行逼近,仿真表明该方法能有效地排除噪声及孤立点对系统逼近的干扰。

This paper modifies the loss function of RCA algorithm, gets a robust convergence by restraining the ill condition in the iteration of competitive agglomeration among the clusters. We use this algorithm for function approximation with TSK model. The simulation shows that the algorithm has good performance of restraining the interference of outliers.