Abstract
Afuzzy rule-based system, widely used in the areas of control and pattern recognition, is mostly structured into inference mechanism and fuzzy rules expressed in terms of fuzzy sets. In this paper, we provide a general framework for fuzzy inference, which consists of a learning part and an inferring part. The learning part analyzes histograms of learning data to generate fuzzy sets and correlation matrix. The inferring part processes test data and draws conclusions with the model built up in the learning part. To confirm the effectiveness, we applied the suggested system to a pattern recognition problem with the iris data. We believe that the suggested system can also be used in control problems.
Get full access to this article
View all access options for this article.