The Improvement Of Apriori Algorithm On The Basis Of Rough-Set Theory

Association rule is one of the methods of data mining. By analyzing training data set, we can find potential and valuable knowledge according to association rule. In the area of business, association rule mining is widely used. As we all known, the association of"napkin and beer"is found by using association rule mining.At the beginning, this paper introduces the conception of data mining and data mining alorgrithm. In association rule mining, Apriori algorithm occupies very important status. Apriori algorithm filters and analyzes items in the database, and finds strong association of the frequent items according to minimum support, then calculates the confident value of the items. However, Apriori algorithm will produce a mount of candidate items and scan the database so many times, which will reduce the efficiency of the algorithm.And then the paper introduces the conception of Rough-set methods. Rough-set is a kind of theory to analyze and deduce data. Rough-set can efficiently deal with the data and information in complex system. And it is a kind of tools to deal with fuzzy and imprecise problems.So, we propose a new method, AprioriRS alogrithm to improve the effect of Apriori algorithm by using the method of Rough-set. AprioriRS alogrithm pre-deals the data in the database by using items division. Thus, we can reduce the possibility of the candidate items produced. In the process of items filtration, we use the method of knowledge reduction which is one of the methods in Rough-set to reduce the items. This can avoid the operation to scan the database repeatedly.And then, we have done some experiment about this. By mining training data, we compare the results of both the algorithm improved and Apriori, and get an ideal result.