期刊名称：Lecture Notes in Engineering and Computer Science
印刷版ISSN：2078-0958
电子版ISSN：2078-0966
出版年度：2019
卷号：2241
页码：409-414
出版社：Newswood and International Association of Engineers
摘要：Fuzzy modeling has been extensively studied.
Among them, it has been shown that fuzzy modeling methods
using vector quantization (VQ) and steepest descent method
(SDM) are effective in terms of the number of rules (parameters).
The methods firstly determine the initial parameter values
of fuzzy rules by using VQ with learning data, and then tune the
parameters by using SDM. On the other hand, Neural Gas (NG)
is known as a novel approach of VQ and NG with local linear
mappings (LLMs) has been applied to a time-series prediction
problem. In the application, the predicted value in each of
subregions is approximated by using a corresponding linear
mapping. It has been demonstrated that, compared with Kmeans
with RBF, NG with LLMs is advantageous in terms of
the accuracy and the number of subregions. The idea of NG
with LLMs has been applied to fuzzy modeling with TS fuzzy
model, and its effectiveness has been demonstrated. However,
the effectiveness of this approach has not been been confirmed
for simpler fuzzy model such as simplified fuzzy model. This
paper proposes to apply the concept of NG with LLMs to fuzzy
modeling with simplified fuzzy inference model. The proposed
method firstly determines the initial parameter values of fuzzy
rules including the weights in the consequent part by using
NG and supervised learning, and then tunes the parameters by
using SDM. The effectiveness of the proposed method with simplified
fuzzy modeling is demonstrated in numerical simulations
of function approximation and classification problems.
关键词：Simplified Fuzzy Inference Model; Vector
Quantization; Neural Gas; Local Linear Mapping
