Application Of Data Mining Technology In The Tax Source Monitoring Based On Clustering

With the development of computer, communication and artificial intelligence technology, social informatization is deepening. Business information, medical information, government information......all the information filled in people’s lives, so that we live in a society that has become the information society. It profoundly affects people’s work, study and life. On the one hand it brought about various facilities. On the other hand, it also produced huge amounts of data. The issue of "expansion of data but poor knowledge" has become increasingly prominent. In the era of knowledge economy, knowledge creates wealth. The question of how to grasp the new knowledge has become the focus of competition. Therefore, mining from previously unknown knowledge hidden behind the mass of information, has become a new hotspot of research and application.As for the country, for the people collecting tax revenue, the department also faces the serious problem that tax-related data isn’t depth of processing and utilization, wasting a lot of valuable information. With the continuous progress of the "Golden Taxation Project (third phase)", the tax department data rise further as data centralization from provincial to national, tax collection databases accumulate a large number of internal and external data. But the current processing and analysis of these data for the tax department only in the application of query, summary, simple comparison, lacking of deep mining of contains information from data.This paper tries to introduce data mining technology in the tax management, to improve scientificalness in tax policy and quality of tax service, and focuses on the application of clustering algorithm in tax source monitoring. The tax source monitoring as a basis for tax management activities, goes through the tax declaration, tax service and tax audit process. Grasp the sources of information completely, and implement strict control measures, can effectively guarantee the tax stipulated in the law and the realization of national macro policy. A clustering algorithm in data mining has been widely applied in many fields such as medical, business, biology. The clustering algorithm applied to the tax source management, can effectively improve the scientificalness of the tax classification, and can help the tax management personnel constitute different tax source management decision. It has the important practical significance for the tax department work efficiency and management level.This paper summarizes the domestic and foreign research and application experience, the technology of data mining, clustering algorithm, K-means clustering analysis and tax source monitoring and other aspects of knowledge. Then we put the two aspects of knowledge together, analyzing the necessity and feasibility of clustering algorithm that should be fully used in the tax source monitoring, and designing the clustering algorithm in application of monitoring tax source model. Then we use SPSS Clementine12.0software to cluster166taxpayers in strict accordance with the CRISP-DM "cross industry standard process for data mining" model, which is divided into three categories of the general source, the key source and the high quality source. In the empirical analysis, we got data from the big wisdom soft according to our index system. Then we put forward some suggestions on various sources of tax supervision. Finally, based on the summary of clustering algorithm used in the management of tax resources, we put forward opinions and suggestions related in tax source monitoring for the application of clustering technology.