A new algorithm of association rules mining

Abstract

To reduce the number of candidate itemsets and the times of scanning database, and to fast generate candidate itemsets and compute support, this paper proposes an algorithm of association rules mining based on attribute vector, which is suitable for mining any frequent itemsets. The algorithm generates candidate itemsets by computing nonvoid proper subset of attributes items, it uses ascending value and descending value to compute nonvoid proper subset of the weights of attributes items, the method may be used to reduce the number of candidate itemsets to improve efficiency of generating candidate itemsets. And the algorithm gains support by computing attribute vector module, the method may be used to reduce the time of scanning database, and so the algorithm only need scan once database to search all frequent itemsets. The experiment indicates that the efficiency of the algorithm is faster and more efficient than presented algorithms of congener association rules mining.