Association rule mining in large databases is one of the most significant contributions from the database community in KDD (Knowledge Discovery in Databases). During the last few years, fuzzy set theory has been integrated into many conventional association rule mining algorithms. However, these studies on fuzzy association rule mining often assume that the fuzzy sets and fuzzy membership functions are known and given. In large databases, it is almost impossible for users to provide all fuzzy membership functions of fuzzy sets for the attributes involved. This paper explores an algorithm to efficiently obtain fuzzy sets and fuzzy membership functions from large databases. The proposed method is based on an integrated K-means clustering algorithm. The preliminary results appear promising.
展开▼