Application Research Of Clustering Analysis On Management Of Tax Resources

In recent years, the research of data mining has developed rapidly and the construction of informationization in tax system has also received great progress. Lots of management systems have accumulated plenty of data. But, the application of data mining in tax system is still in the early stage, how to apply the technology of data mining in tax system has important meaning both in theory and practice.This dissertation uses data mining technology to deal with tax source sorting management. Detailed work is as follows:(1) Briefly introduced the general situation of construction of informationization and application of data mining in tax system.(2) Briefly introduced the production,development and application of Knowledge Discovery in Database (KDD) and Date Mining (DM), researched the clustering algorithm and the standard of measurement in clustering analysis.(3) To solve the clustering problem of data sets with data missing, we modified the original K-means algorithm through changing the method of measuring distance and generating new center, and then we presented an improved clustering analysis algorithm for data sets with data missing. We did some experiment using standard UCI data with parts of data removed. The result shows that this algorithm is feasible and effective.(4) To solve the problem of tax source sorting, we chose the tax register data and the detailed tax payment data of one city in AnHui province as data source, and then analyzed them using the technology of clustering analysis. Before analyzing, we disposed the data by data-choosing,data-locating and data-extracting to measure up our request first, then loaded them into the mining-base, standardized the loaded data, finally, after analyzing, we got the information of tax source sorting with highly instruction meaning to tax source sorting management.