Loading...
Fast adaptive fuzzy autoregressive model
Author(s)
Bunyarodol, Decha
Date Issued
December 1, 1999
Type
Conference Paper
Abstract
In this paper, a fuzzy autoregressive model with a multiple criteria adaptive algorithm is proposed. The proposed algorithm consists of two independent criteria for adjusting the fuzzy rules and some related parameters. First, the proposed algorithm adapts the existent fuzzy rules by gradient descent method based on predicting error criteria. In addition to rule adaptation, the learning rate of each rule adaptation can also be adjusted based on matching of rule confidence of each fuzzy rule. With the combination of the two independent criteria for adaptation, the proposed algorithm can increase the speed of learning.
Citation
Intelligent Engineering Systems Through Artificial Neural Networks, 9, 581-586, 1999