Tax Assessments Based On Data Mining System Design And Implementation

Taxpayer Compliance Management has become an important part of taxation management in China.Unfortunately,constructing with the traditional technology,the Taxpayer Compliance Management information system meets a lot of problems,which are considered to be resolved by the data mining methodology.Aiming at satisfying the TCMIS's requirements,we investigated in depth all kinds of Decision Tree Algorithms,especially the SLIQ Algorithm,which is capable of handling huge volume of data with high efficiency.But SLIQ Algorithm is hard to implement, and can not utilize the data stored in DBMS directly,and the calculation of gini index is superabundant during the construction of decision tree.Therefore,considering that in TCMIS,the huge volume data which is stored in DBMS,based the concept and theory of Gini Index of SLIQ Algorithm,we design the QLIQ Algorithm,which can build decision tree on VLDB data.The QLIQ Algorithm not only takes the efficiency into account,but also is designed to be easy to be implemented. Especially,it can take full advantage of DBMS to process high volume data.At last we ameliorated an existing TCMIS with QLIQ Algorithm,and developed the proto type of evaluating object filtration module to examine this algorithm.The experiment indicates that QLIQ Algorithm can resolve the problems existing in TCMIS to meet the business demand,and it is efficient and can be easily implemented.