| Summary: | 碩士 === 國立臺灣科技大學 === 電子工程系 === 89 === In recent years, many methods have been proposed to generate fuzzy rules from training data. In this thesis, we present a new algorithm (FRG) to generate weighted fuzzy rules from a set of training data, where the attributes appearing in the antecedent parts of the generated fuzzy rules may have different weights. We apply the generated weighted fuzzy rules to deal with the “Saturday Morning Problem”, where the proposed FRG algorithm can get a higher average classification accuracy rate and generate less fuzzy rules than the existing methods. Then, based on the genetic algorithm, we propose a new method consists of the FRG algorithm to tune the weights of the attributes appearing in the generated fuzzy rules for generating weighted fuzzy rules. We also apply the generated weighted fuzzy rules to deal with the Iris data classification problem. The proposed method can obtain a higher average classification accuracy rate and generate less fuzzy rules than the existing methods.
|
|---|
