Research On Data Template Matching And Application Of Social Security Audit

The social security relates to the vital interests of the people, so its supervision and examination are the top priority of audit work. Because of the diversity and heterogeneity of social security information systems, it often encounters data items inconsistency when the data in social security database imports to audit database. Currently, the method to solve the problem is to establish the corresponding relationship between the source template and target template manually. Although the accuracy of mapping between two templates established manually is high, it always takes much time and effort. So it has become a serious problem to implement semi-automatic/automatic matching of data template in social security audit.This thesis studies the pattern matching technology existing both at home and abroad, a data template matching model which based on semantic is proposed to deal with two major problems in template matching process, the one is that matching space is too large to determine the scope of matching, the other is that it is difficult to found matching relations due to irregular name. The research of the model consists of two parts, pattern categorization and matching generation. The pattern categorization includes categories identified of data template in social security audit, classified information extraction and processing, the implemen- tation of pattern categorization based on Na?ve-Bayes and HowNet, and the combination of categorization results. Matching generation consists of pattern matching generation based on the similarity flooding algorithm and the method to choose candidates.Finally, the model of data template matching in social security audit is checked by experiments, and the experimental results are analyzed to verify the correctness and efficiency of the model.