Research On Association Rules In Data Mining And Its Application

Association rule mining is one of the most important parts in data mining. The typical algorithm of association rule is Apriori which was put forward by RAgrawal. The proposed algorithm scans the whole data base during each loop in calculating support of the candidate itemset. However, the efficiency of mining is low. Therefore more effective algorithms must be developed.This thesis explores association rule algorithm in data mining, and analyzes the classical algorithm Apriori in detail. The procedure of the proposed algorithm is presented and the deficiency is proposed. An efficient association rule algorithm NEA is developed in terms of the defect of the proposed algorithm by using the results from sets of L_k and C_k to filter out the database, therefore the number of records searched in candidate sets from the database is reduced, and the efficiency of the algorithm is enhanced. The efficient updated algorithms Minsupchange and Minconfchange are presented when the values of minsup and minconf increase or decrease. And the comparision with Apriori is made. Finally, the improved algorithm NEA is applied in the evaluation of teaching. In order to find out the relationship between teaching and the qualification of the teachers, data mining is performed, thus the teaching efficiency and quality are improved.