Computer-aided Design Of Case Selection And Nlidb Achieve

Tax revenue is the lifeblood of a country, tax audit is the most important work of tax collection and administration. Case selection is the first link of tax audit, whose aim is to find out the delinquent taxpayers according to the scientific case selection criteria using various kinds of taxpayer's information concerning taxation. Computer applied to case selection can produce great positive effects and can help to solve the administration problem of human resource inadequacy.Case selection criteria are the basic of case selection, which define some criterions, by which a taxpayer can be evaluated if he is the delinquent one. In the previous computer aided case selection systems, the person who takes charge of maintaining the case selection criteria should master not only tax knowledge but also computer technology, which brings so much difficulties to those inspectors of taxes. Natural Language Interface to a Databases (NLIDBs) are provided in this thesis to solve this problem.NLIDBs provide users with a way to access information stored in databases directly in natural language. It communicates the mode of understanding information in the real world with the one in the machinery world. NLIDBs involves many kinds of subjects, such as AI, NLP, DB, HCI, etc.The computer aided tax system described in this thesis is a semantic grammar system, with the semantic dependency tree as the intermediate language. The semantic dependency tree is based on dependency grammar and database semantic. Having good corresponding relationship with SQL structure, the semantic dependency tree can be translated to SQL and paraphrasing to natural language easily. Thus we have put the database semantic into the natural language process, which can solve some problems in the previous approaches to NLIDBs.The computer aided tax system has passed the test in the real environment.