The Research On Expansion Of DM2 Platform And Application Of DM2 Platform To Data Mining In Railway Freights Tickets Data

Database technology has been extensively on the popularization and application from the early 1980s. In recent years along with the rapid growth of the amount of data and the growing popularity of data warehousing and Web data sources, the main problems that people are facing is not lack of useful information, but how to make effective use of enormous data. Faced with this challenge, data mining technologies have emerged and have been widely applied in all walks of life. Association Rules mining is the most active one of directions of research on data mining.DM2 platform is a data mining platform designed and was developing by us. Now DM2 platform has developed the function components for instances classification and association rules mining, and has implemented a variety of data mining algorithms, such as the ID3, Naive Bayes, FP-Growth and Closet. In order to meet the demands of the different projects for data mining, there are many parts need to improve and expand in the DM2 platform. Firstly, the capabilities that DM2 platform interact with database need to be enhanced. Secondly, the algorithms implemented in DM2 platform are very limited, so to implement more data mining algorithm in DM2 platform is first imperative because an algorithm is precisely the essence of data mining system. Thirdly, DM2 platform does not have interaction functions with users through the interface. Aim at these issues, this paper expand DM2 platform from how to rich algorithm library and how to implement user interface.Firstly, this paper add data mining algorithms to DM2 platform, it implemented a classic Apriori algorithm, presented and implemented an improved Apriori algorithm. This algorithm adopted linear data storage structure and vertical data structure of database, solved the bottelneck of the classic Apriori algorithm, improved the performance of DM2 platform in a certain extent.Secondly, this paper has been further strengthened the ability that DM2 platform interact with database, implemented the storage from rule set to the database and make the DM2 platform can filter, sort and group rules.Finally, this paper implemented the user interface of DM2 platform. It achieved dynamic interaction with users by browser-based technology and JSP technology and gained mining results intuitively.The expanded DM2 platform has extremely capacity to handle large data sets, excellent ability to interact with database and human visual interfaces of data mining.Taking railway freight tickets recording sample of Zhengzhou Railway Bureau in 2004 as experimental data, this paper established railway freight data mining system based on extended DM2 platform. Experimental results show that the system can mining valuable association rules and has reliable performances.In the last part of this dissertation, we summarized our work on DM2 platform and analized the improvements to be done in the future.