An Approach of Improved Traversal Merging of Transaction Data for Faster Apriori Algorithm

Abstract

In order to improve the operational efficiency of the Apriori algorithm in the data preprocessing stage of large-scale data and achieve overall optimization of the Apriori project, a fast traversal merge pre-processing method is proposed by integrating an adaptive association mining threshold determination method. Firstly, the proposed fast traversal merging method is analyzed and compared with two benchmark algorithms, and the experimental results show that the running time of the fast traversal merging method is much lower than that of the two benchmark methods; secondly, according to the central limit theorem, a data adaptive support threshold setting method is proposed, which can avoid the subjectivity of the minimum support threshold setting in association mining; finally, the two proposed algorithms are applied to Apriori and the results show that the application of the proposed improved method for association mining gives significantly better results than association mining under the better processing of the benchmark algorithm, and thus can significantly improve the efficiency of solving the shopping basket problem.