Research And Application Of On-Line Analytical Mining In Government Statistic

The duty to statistics departments includes building statistical database for storing and managing the statistical data centrally, collecting, gathering and analyzing data etc. However, the rapid development of the information progress in statistical work has made the traditional statistical tools obsolete.The character of statistical database just accords with the internal demand of OLAP (On-line Analytical Processing), because of its multidimensional data structure. It can provide statisticians with valuable information from statistical data, by mining the multidimensional data sets produced by On-line Analytical Technology. Therefore, this thesis will show a new way of combining OLAP with data mining applied to the socio-economic statistical work.The application of On-line Analytical Mining in government statistic has been studied deep in this thesis. According to a framework of OLAM realization towards statistics, multidimensional data models for various analyses are researched and established. The basic statistical data have been gathered by some general aggregation algorithms based on OLAP. So multidimensional data sets have been obtained, including the statistical grouping and statistical measurement. In addition, the aggregation algorithms don't depend on special API provided by any RDBMS. So the algorithms can achieve cross-platform of database environment and adapt to various database systems. Meanwhile, due to the typical time series nature of the statistical data, the trends, comparability rule and association rule about time series within the produced multidimensional data sets will be mined in the thesis.Finally, it is introduced how the research has been applied successfully to a statistical information management and application system that is specially oriented the area and county statistics departments. At the present time, the system has served Minhang area statistics bureau in Shanghai and the phasic production has been gained. The integration of statistical data resource has been made by the statistical information management system, avoiding the waste of abundant statistical data. As a.flexible, standard and general statistical data management platform, it provides friendly interactive interfaces and rich data mining representation for the statisticians. This system meets the demands of the statistical work, and improves the efficiency of statistical work. So the policy of the government economy can be shaped.