Application Study Of Data Mining In The Highway Toll System

With wide application of the Freeway Toll System (FTS), large amounts of toll data are emerging. Accordingly, acquiring useful knowledge through data analysis and processing has become a key problem in the field of data processing. Data mining is such a novel technology for solving this problem. It can deal with large amounts of historical data and then dig out hidden information, pattern or tendency from them. Therefore, data mining has been considered as a promising technology.On the basis of practical investigation and primitive data analysis, this thesis employs data preprocessing technologies such as data extracting, transforming and loading so as to build a subject-based data warehouse of FTS. First, for an overweight subject, the overweight vehicles are classified by the decision tree algorithm and the Bayesian algorithm, respectively. Second, for a toll subject, the toll amount is predicted by the linear regression algorithm. Finally, the software integrated development environment Visual Studio 2005 combined with the data analysis tool Microsoft SQL Server 2005 are used to process primitive data and get final results.Our system firstly integrates toll data with the data warehouse technology and processes the data in the warehouse with the data mining technology, then obtains some useful conclusions which may provide effective decision support for highway managers.