Research On Association Rules Analysis Technology In The Audit Field

Association rules is an important research topic of data mining, the purpose of mining is found the implicit association between the item in the database.Based on the study the positive and negative association rules, and applied its mining algorithm to the field of social security auditing, access to the factor of audit characteristics hidden in the audit data,mining association rules with negative items in the audit data, analysis of its internal audit significance and will contribute to the improvement of audit methodology.This thesis summarizes research results in related fields at home and abroad, introduced the basic theory of data mining and association rules, including the basic process of data mining, key technology, classification of association rules and the basic concept of negative association rules, and discussion the accuracy questions of association rules combined with a variety of knowledge. For the pretreatment problem of structured data, in summary and in-depth study based on the data cleaning and conversion issues were discussed; For the problem of mining the factor of audit characteristics in the audit data, MHP algorithm based on a hash table been proposed, thus combination the frequent itemset mining with the process discovery of the factor of audit characteristics, and applied to social security audit data set, analyzed the performance of and the results of the algorithm; Analysis of a typical the positive and negative association rule mining algorithm:PNARC algorithm, for the inadequacies of the algorithm, a PNAR_M algorithm base on minimum correlation and double confidence been proposed,and performance of the algorithm was experimentally verified; These two algorithms are applied to the audit data set, mining association rules with negative items between the audit data,and interpretation of the mining results, ultimately elaborate on its significance in the audit.