Research on Pruning Techniques of Mining Weighted Sequential Patterns

Abstract

The research of sequential patterns mining is very hot, and a variety of classical sequential patterns mining algorithms have emerged, and some data mining tools have been developed for free study and use. There are very few references in this new field, and the time and space cost of mining are large. This paper presents a new pruning technique. In this paper, we introduce minimum support and the k-weighted expectation pruning strategies in the weighted negative sequential patterns mining algorithm. The data set used in this paper is provided free of charge by UCI's official website, using the improved k-WNGSP mining algorithm and the existing WNGSP algorithm. Under the same condition, it is found that the number of negative sequences that can be excavated by the k-WNGSP algorithm. We can say that the number of the sequences is increased and the time consumed is shorter. Experiments show that the algorithm is effective and obtains the ideal experimental results.