Research And Application Of The Rough Set Method In The Tax Audit

Tax audit cased-selection is the knowledge and judgement of tax bearer credit, the research of that will exert direct affection to the reasonableness of tax audit plan, and it is a very important area in the tax research. The traditional method is set up on the experience of expert or the simple statistics model, it's difficult to judge the high non linearity relationship between the tax declare status of tax bearer and the tax dodging, so, that can not meet the requirement. So, it's necessary that use the data analysis technology to deduce the reliability of the tax. In this paper, the main content below are included.The application research of how the rough set method is used in the tax audit cases-selection. Based on the introduction of rough set theory and method, a simple algorithm is given because of the relativity and redundancy among the specification of cases-selection.â– A model is set up based on the rough set- learning vector quantization neural network. The model combined the rough set ability and the strong adaptability of neural network on the non linearity problems. With the key specification of rough set, it not only can decrease-the amount of the necessary data for the study of neural network, but also can increase the intelligibility of potential knowledge, and can improve the ability and category resolution.The research has been used in the CTAIS province level data integration platform in Gansu State Tax Bureau, and the prototype system of tax audit cases-selection was executed based on the rough set- learning vector quantization neural network. A higher precision of cases-selection is achieved during the field application. Now, the system has been used and the result is good.