Research On The Railway Passenger Ticket Data Minning

With the continuous development of computer science, a lot of raw data has been collected continuously and stored in the computer. The low ability of data processing resulted in the current status of rich information but poor knowledge. Data Mining - using non-trivial way to find useful knowledge from a large amount of data, is the response to this request and quickly developed as a science.Railway passenger tickets system contains a wealth of information. It is an important problem that how to mine useful knowledge from the mass of information. Aiming at the characters of train tickets and applying the technology of data minning in tickets analysis, we educe information of the passenger stream and influence rules of passenger tickets' characters to the action of buying tickets. So that we can use them to direct the organization of rail passenger transport and serve the marketing decision of rail passenger transport.The SQL Server 2000 is chosen as the development tool of data warehouse and data mining. The basic principles of data mining technique, visual data mining technique, OLAP and data warehouse are presented. Decision tree and cluster analysis in the SQL Server 2000 are highlighted. After the preliminary analysis to the data in the railway passenger tickets system is carried on, the integration and conversion to several databases involved are carried on. The detailed design process and step of the data warehouse for decision making are provided. As the sample data, the tickets data is collected from Beijing-Kowloon line in Nanchang bureau with downward direction in April 2008. A star framework cube is established. Characteristics of samples data are analyzed. The data mining to the seat, leaving time and train variety is carried on. Last result analysis is described.