A Hybrid Learning Algorithm for Fuzzy Neural Networks

摘要：

We propose a novel hybrid learning algorithm for fuzzy neural networks. The algorithm consists of the gradient descent method and a recursive SVD-based least squares estimator, which are used to refine the premise and consequent parameters, respectively. The advantages of our method are that the consequent parameters are updated optimally and that the search space of backpropagation for premise parameters is greatly reduced. As a result, our algorithm converges more quickly and produces smaller errors than the pure gradient descent method.

作者: Chen-Sen Ouyang Shie-Jue Lee

作者单位: Department of Electrical Engineering National Sun Yat-Sen University Kaohsiung 804, Taiwan

会议类型: 国际会议

会议名称: 8th International Conference on Neural Information Processing(ICONIP 2001)(第八届国际神经信息处理大会)

会议地点: 上海

会议语种:英文

页码: 349-354

在线出版日期: 2001-11-14（万方平台首次上网日期，不代表论文的发表时间）