Research On Sampling Of Tax-checking

The development of national information system and network is changing the structure and functional mode of economy,society and culture. The information technology already became the basement of economical development and it brings new challenge and demand for taxation administration. National taxation administration advocates that reinforce levy and mangment, build up leak, punish corruption and hand over owing tax. This research uses information technology in Tax-checking and makes use of data mining to distill some useful information that people do not know in advance from the great deal of data that belongs to tax administration. The method can help to improve the work efficiency and management level and powerful support of technology. Based on the national tax auditing practice, this paper analyzes the use of Self-Organizing Feature Map in value added Tax-checking sampling to detect taxpayer non-compliance. Firstly, the meaning of Tax-checking sampling is analyzed and the current research situation is summarized. Secondly, this paper analyzes the meaning and functions, operation flow, basic procedure and points out the information resource and main methods of current Tax-checking sampling and analyzes these questions and reasons. With the help of knowledge of information economics, financial management, statistics and pattern recognition with Artificial Intelligence, the theoretical basement of the method is put forward.Thirdly, in this rearch, we discussed the application of discriminant analysis, and created an indicator-selecting model of value added Tax-checking sampling. We also give an analysis model of value added Tax-checking sampling based on the Self-Organizing Feature Map. Finally, with an instance, we use discriminant analysis and SOM to discuss the indicator selection and clustering of value added Tax-checking sampling. According to the practical data from enterprises, we testified that this method is effective.