MULTI-OBJECTIVE GENETIC-FUZZY DATA MINING
Chun Hao Chen, Tzung‐Pei Hong, Shih-Pang Tseng, Lien Chin Chen · 2011
Abstract. Many approaches have been proposed for mining fuzzy association rules. The membership functions, which critically in uence the nal mining results, are dif-cult to dene. In general, multiple criteria are considered when dening membership functions. In this paper, a multi-objective genetic-fuzzy mining algorithm is proposed for extracting membership functions and association rules from quantitative transactions. Two objective functions are used to nd the Pareto front. The rst one is the suitability of membership functions. It consists of the coverage factor and the overlap factor and is used to avoid two unsuitable types of membership function. The second one is the total number of large 1-itemsets from a given set of minimum support values. Experimental results show the effectiveness of the proposed approach in nding the Pareto-front mem-bership functions.