Research On Sequential Pattern Mining

The sequential pattern mining is one of the most important studies on data mining. After the sequential pattern mining is put forward,it produces a lot of classic algorithms, for example Apriori All algorithm,GSP algorithm,SPADE algorithm,Free Span algorithm and Prefix Span algorithm.In this paper,the GSP algorithm is improved by the time consuming problem,which can reduce the time consumption of the GSP algorithm.Firstly,this paper introduces the researchs' background and the content of sequential pattern mining was introduced.Then,it describes the basic concepts and some kinds of typical algorithm in sequential pattern mining.The focus of this article is from three aspects to improve the efficacy of GSP algorithm.First of all,after analyzing the data structure of sequential patterns' mining,the time index is introduced in the 0-N data structure,which construct the 0-N data structure with it.Secondly,in order to get the complete sequence pattern,the time index worked as a division goal to distinguish the sequence databases' level.In the end,the transaction compression technology is introduced to improve the algorithm's scaning efficiency.Finally,this paper summarizes the content of this paper and prospects the future.