Data Mining In Auto After-sales Service The Application And Research

As the landmark service providers in the auto after-sales service industry, the 4S stores now used a variety of application software systems for business process refinement and standard of service, on the same time, a large amount of historical data was stored. Though the corporate's decision-making management has been aware of the value of those data, but not too good in terms of use. The 4S stores in the country start relatively late in itself, they are not yet perfect in all aspects though development so far, so there are much room left to improve service quality. In this particular environment, the application of data mining from historical data will make a great contribution to further development to the enterprise.The task of this thesis is extracting the vehicle repair and maintenance related data on the basis of existing database of the 4S stores on the basis of existing database to establish the data warehouse system for auto after-sales service industry, which store the extracted data after the process of data cleansing and ETL, and then doing some rearch on data mining application by the method of association rules and statistical methods. There are two aspects, first is doing correlation analysis to customers'vehicles of the 4S stores, researching vehicle maintenance history data, analyzing customer's maintenance cycle and and the projects of repair and maintenance under ordinary circumstances, roughly understanding the driving habits of customers and the situation of roads which vehicles usually use, supporting Customizing personalized service for customers with data. Second is doing rationality analysis for the allocation of parking spaces of the enterprise's transparent workshops and job schedule of maintenance team, improving service quality and efficiency of the enterprise by reasonablly scheduling maintenance team in the case of maximizing the use of corporate resources.This classic Apriori association rules algorithm for the experimental study, the algorithm generated a large number of frequent item sets and repeated access to data in these two shortcomings of things caused by time-consuming problem, based on FP-Tree based technology improved Apriori algorithm FP-Tree algorithm does not use the candidate items, but also directly compress the database into a number of frequent patterns, and then association rules generated by the tree, the efficiency has improved significantly.In this paper, data preprocessing before the data mining and data mining the results of the analysis and display have also been some research on features for enterprise data, data conversion algorithm to adapt to the application, the basic achieve better results. Reports on the form of mining results to show to the customer, and to give personalized tips on enterprises to improve service quality high guidance significance.