Research On Tax Data Mining Based On SAS System

With the deepening of the tax information and the full implementation of the tax business application systems, such as comprehensive collection and management system, value added tax management system, the export tax rebate system, and so on. Massive business data is accumulated which is about tax collection and administration and very precious to tax system. So how to effectively utilize these resources to provide scientific and reasonable basis for the decision of leaders, to improve the ability to analyze the source of revenue, or even to strengthen the management level of the tax collection and management, become the most urgent problem to be solved for the current tax system.As a high-level data analysis method, Data Mining is a technology to analyze the large amount of data, so as to reveal the knowledge hidden behind the data. It can help tax authorities from a number of tax data in the analysis of the behavior of tax-payers information, identify the characteristic of escape enterprises, lock inspection objectives, and also reduce audit costs.This paper mainly relies on the data resources of the tax information system in Henan province, based on the current international advanced data mining tool SAS, using data mining techniques, to explore the data of the provincial tax system. There are two achievements:one is analyzing the data of the go-fled enterprise, using decision tree and logistic regression modeling method to set up the logistic regression model to score the risk management of the enterprise. The other is using mining technology such as cluster analysis, outlier analysis and association rules to analyze the acquiring enterprise in the Henan province, to build the rules library of industry trade, and also positioning false value added tax invoices, which finally set up the model of falsely writing out. At the end of the paper, we focus on the more possible direction of data mining in the tax system.