Tax Assessment And Data Mining

Tax Assessment closely influences all businesses of tax at all levels. Good tax assessment resultscan be used as a macroeconomic analysis, tax basis for structural analysis and analysis of the industryboom. It also guides the tax risk governance, management and audit options of key sources offacilitated. It is a core mission of decision system of taxation.On the other hand, the database stored in the tax information system has far exceeded thecapabilities of analysis and processing, the data is rarely accessed and employed, decision makerslack the right tools to get valuable information from massive data to guide decision making.The fundamental objective of data mining is to find particular patterns in the data. Therefore,Data Mining should be introduced to Tax assessment urgently, which has formed a consensus in thetax system. In the early stages of the current application, tax staff always can not grasp basic methodsfor analysis goals, not sure how to select mining techniques or algorithms. This is mainly blocked bybarriers between technical areas and business areas. No getting rid of the barriers, no breakthrough inTax Assessment.In order to adapt to the realities of tax assessment needs, article briefly summarizes DataPreprocessing, describes several processes in detail, including Attributes Reduction, Cluster Analysis,and Anomaly Analysis, and proposes a series of novel ideas and methods. By actual cases of taxevalution, it shows the rationality of ideas, as well as the effectiveness of methods.The article brings forward the following new knowledge points:①Property Weights Based onDeburring Algorithm;②Attributes Reduction Algorithm based on Boolean Matrix Filter;③Directional Cluster Analysis and Weighted Distance;④Anomaly Analysis of type I and type II.