Research And Application For Parallel Mining Algorithm Of Association Rules In The Social Security Audit

Association Rules (Association Rules) found interesting or relevant connection between the set (Itemset) of a large number data was an important part of research in the field of data mining. Association rules applied to social security data will be found lots of the potential knowledge of computer-assisted audit social security data. Faced with the huge social security data, Parallel Mining Algorithm of Association Rules is an effective method.Firstly, the thesis was explained the significance of the social security audit and analyzed the role of the Association Rules of Data Mining in social security audit. Secondly, the development of data mining, research, applications and the basic concepts of association rules (association mode), commonly used algorithms and expansion algorithm and the parallel association rules algorithms and results were introduced. Finally, at basis of concretely analyzed the characteristics of the social security data, association rules parallel sampling algorithm was introduced to audit of social security data.The four-node association rules parallel sampling algorithm designed in the thesis was firstly applied to the field of social security audit. The association rules parallel sampling algorithm was divided into two steps: First, each node independently using classical algorithm APRIORI of association rules Mined respective data. Second, Will be designed to each node of mining results transmitted to the root node, gathered the results of the excavation node and output of the association rules. Based on the experiments, after compared the result of sequence of time division, the results of regional division and the results of random strategy division, the sequence of time division was applied because of it's higher precision. According to the characteristics of data, data extraction, data conversion and cleansing and data loading was mainly disposed.In this thesis, association rules parallel sampling algorithm combined with the social security audit, using the technology of association rules to audit the social security data obtains the better results. At the same time, the efficiency of data mining improved greatly.