Research and application of association rule mining algorithm based on multidimensional sets

In this paper, we describe the basic concepts of multidimensional sets and multidimensional association rules, and propose a improved algorithm of association rule mining based on multidimensional sets. This algorithm can find out the maximal frequent item sets of each dimensional subset, and at the same time pruning the database, this can substantially reduce the workload of the subsequent mining and lead to efficient processing. Appling the algorithm to students´ synthesis marks evaluation system, analyzing the related factors which influence the students´ synthesis marks, for educators to improve teaching methods, to improve the quality of talent training.