Research And Implementation Of Sequential Pattern Mining

Data Mining is the procedure of extracting and mining knowledge from large amount of data in database, data warehousing and other information repositor and also is a rapidly emerging research field relevant to artificial intelligence and database system. Discovery of sequential paterns is an important field in data mining research.Sequential pattern mining is to mining patterns that are frequent relative to time or other patterns in sequence database. It is one of the most important tasks of data mining and will have broad application in future such as the analysis of customer behavior in transaction database, Web usage log analysis, the analysis of science experiment procedure, text analysis, DNA analysis and nature disaster prediction et al.Multi-dimensional sequential pattern mining is based on sequential pattern mining and consider some other information, like age,gender or other information in analysis of customers' purchasing behavior. Such patterns combine more information and the value of application is high.On the basis of the analysis of sequential pattern mining algorithm,we convert our focuses on multi-dimensional sequential pattern mining and the specific implementation in the field of application. The main context is as follows:(1) This paper presents a efficient algorithm of mining multi-dimensional sequential pattern. It scan projection database once and record the attribute name and the information of records of all the frequent items when mining MD-patterns,for the following joining. we can get all the MD-patterns just through joining and no need to scan projection database again.(2) The application of anomaly detection using the algorithm of sequential pattern mining. The method of data mining are used to detect abnormal behavior of users fit into intrusion detection system. The behavior database of users are changed into sequence database firstly,then mining frequent sequential patterns from sequence database, compare current patterns with historical patterns and determine whether there is abnormal, the experiment show the feasibility of this method finally.(3) Design and implement a tool of sequential pattern mining. It include several efficient sequential pattern mining algorithms. On the one hand,users can choose suitable algorithms according to their own needs.On the other hand,the tool is a comparisons platform of sequential pattern mining algorithms.Compared with existing tools,it converge several efficient algorithm for sequential pattern mining and provide a variety of options for users..