For Full-Text PDF, please login, if you are a member of IEICE,|
or go to Pay Per View on menu list, if you are a nonmember of IEICE.
An Investigation of Fuzzy Model Using AIC
Shinya FUKUMOTO Hiromi MIYAJIMA Kazuya KISHIDA Yoji NAGASAWA
IEICE TRANSACTIONS on Fundamentals of Electronics, Communications and Computer Sciences
Publication Date: 1997/09/25
Print ISSN: 0916-8508
Type of Manuscript: PAPER
Category: Information Theory and Coding Theory
the destructive method, imformation criterion, AIC,
Full Text: PDF(527.5KB)>>
In this paper we suggest the "goodness" of models using the imformation criterion AIC. The information criterion AIC is a statistic to estimate the badness of models. When we usually make the fuzzy rules, we aim to minimize inference error and the number of rules. But these conditions are the criteria to acquire an optimum rule-model by using the training data. In the general case of fuzzy reasoning, we aim to minimize the inference error for not only given training data, but also unknown data. So we have introduced a new information criterion based on AIC into the appraised criterion for estimating the acquired fuzzy rules. Experimental results are given to show the validity of using AIC.