Research on the calculation method of Apriori algorithm

Abstract

When mining frequent itemsets, the traditional Apriori algorithm uses iterative calculation support level by layer. It needs to scan the database for many times and perform I/O operations frequently. Moreover, it is sensitive to the original data. When the data structure is complex, the number of data items is large, and the density of data elements is high, the problem of low efficiency of Apriori algorithm is obvious. In order to further improve the performance of Apriori algorithm, this paper analyzes the operation principle of the algorithm, and proposes MI_Apriori(Multiset intersection Apriori) algorithm. Based on the equivalent Boolean matrix of transaction records, the intersection operation of multiple sets is carried out, and the maximum frequent itemset is mined through a certain number of iterations. Compared with the traditional Apriori algorithm, the correctness and reliability of the mining results are verified. It shows that the MI_Apriori algorithm has better computational performance, and the improvement effect of the traditional Apriori algorithm is obvious.