ABSTRACT

The most important task in the design of fuzzy classification systems is to find a set of fuzzy rules from training data to deal with a specific classification problem. In this paper, a method to generate fuzzy rules from training data to deal with the data classification problem is presented. Partition method of interval is adopted in current classification based on associations (CBA). But this method cannot reflect the actual distribution of data and there exists the problem of sharp boundary. These type of problems can be approached with fuzzy representation of data. In this paper quantitative attributes are partitioned into several fuzzy sets by fuzzy C-Means algorithm and membership values are generated, and supervised association rule algorithm is used to discover interesting fuzzy association rules, which are used to build classification system. In this paper fuzzy classified association rules are generated and three classifiers namely C4.5, Naïvebayes , and ID3 are used for classification. Experiments are conducted on both primary and secondary data and accuracy of each of the classifiers are discussed with AUC-ROC curves. Quantitative values in databases generate very large number of rules. Using fuzzy linguistic values the generation of rules can be reduced and an objective measure is used further to filter the generated rules and present only the interesting rules.