Research On The User Access Sequence Of Web Usage Mining

User access sequence mining is an important research direction of Web usage mining, which aims at discovering the users' visiting interest and intention of Web sites by mining Web log. With mining results, web site operators can provide users with personalized recommendations and site navigation services. If the frequently access paths are obtained, the reasonable advertisement arrangement can be made to improve site advertising revenue. Site developers can use the mining results to improve site system or structure in order to enhance the efficiency of site visits. In a word, access sequence mining has a very good prospect in various Web site applications, particularly e-commerce sites and portal sites.Access sequence mining includes three stages: data preprocessing, sequences discovery and sequences analysis. The purpose of data preprocessing is to process source data, and the preprocessing results can meet the requirements of mining algorithms. Sequences discovery is mining user's access sequence using sequential mining algorithms, which is divided into two aspects: ordinal sequences mining and sequences patterns mining. The task of sequences analysis stage is finding meaningful knowledge from the mining results.In the thesis, techniques used in data preprocessing have been researched at first, and a preprocessing procedure for access sequence mining was improved. Then in the sequential pattern mining part, a high-efficient sequential pattern mining algorithm named GSP was researched and implemented, algorithm performance testing and analysis of mining results are also provided. Finally, a complete solution utilizing data warehouse platform to mining access ordinal sequences was implemented and verified by using actual commercial site logs. An analysis of the mining results and a patterns compare with GSP was also provided in the end.