Research And Application On The Discovery And Extraction Technology Of Audit Monitoring Indicator

Data mining is one of the most important technologie of the field of machine learning,and is an important process of knowledge discovery. Its purpose is extract and mininginformation and knowledge from the massive, incomplete, noise, fuzzy, random data, which isnot known in advance to people, but is potentially useful. The existing computer auditmethods are scattered, which are processed and classed manually, and audit monitoringindicator are found and extracted manually, which make audit experience has a low sharingdegree and bring a lot of inconvenience to audit workers.This thesis is based on the summary of the data mining technology research status andresearch results at home and abroad, and introduces the data mining technology and thetheory knowledge, especially the technology of the association rules analysis and clusteringanalysis For the existing audit thesaurus’ deficiency, use the unknown word identificationalgorithm, gives the dynamic renewal process of the social security audit database. And onthis basis, take the sets of the social insurance audit as data sets, for two defects of theK-means algorithm of the clustering algorithm, give the concept of the minimal rule coveringset of the association rule analysis technique for the first time, and propose the based-theminimal rule covering set of K-means improved algorithm, improve the clustering resultseffectively, and successfully applied to the process of the audit monitoring indexes’ detectionand extraction, and further have the study of the method of the audit monitoring indexes’detection and extraction The audit monitoring index detection and extraction systems alsoemerge as the times requres, and has laid a good foundation for the latter part of the auditwork.