Model And Analysis Of Algorithms Based On Customer Behavior Railroad Web Log Mining

The application of Railway Online Ticketing System has produced a large mounts of data in users’ information and operation information. These information contain potential customer behavior mode. Through the analysis of these data, we can get not only the influential factors of passengers’ purchasing but also the user’s preferences. These results of analysis provide an important decision support to the work plan of railway transport sectors. This paper use web logs and railway passenger ticket data as the foundation of analysis, and use Artificial Neural Networks as the modeling approach. This paper study on the problem of railway passenger clustering and online sales prediction, and the main works is stated as follows:Firstly, Railway Customer Behavior Analysis Indexes System based on Customer profiles, purchase behavior and using preferences is builded. This analysis indexes system contains ten Secondary analysis of indicators. In the context of the present situation of railway e-commerce and the research needs, the specific definitions and the measurement methods of these indicators are given.Secondly, Railway Customer Segmentation Model based on SOM-Kmeans hybrid algorithm is proposed. This model use web logs and railway passenger ticket data as the foundation of analysis, and consider the definition and metrics of the relevant indicators. Then the using preferences matrixes in railway online ticketing system is established, and use the matrix as the segmentation base to divide the railway customer groups.Thirdly, Railway Online Sale Prediction Model based on improved BP neural network is proposed. This model not only thinks of the number of main customer groups, but also considers the holiday influence as impact factors. And then forecast the sale volume of online ticketing system.Finally, Railway Customer Behavior Analysis Prototype System base on the above models is designed and realized. Using the technologies of VC++and SQL server2005to develop the prototype system. The effectiveness and availability of prototype system is provided by the analysis results.The research findings of this paper can supply a reference to the railway passenger transport sector in the field of railway passenger behavior analysis and online sale volume prediction.