Research On Data Integration And Key Technology Oriented To National Audit

Network audit embodies advantage of resolving audit challenges followed by rapidexpansion of the amount of data, and greatly reduces the cost of the audit, so, it got rapidpromotion in a short period of time, and gradually replaced the traditional means of audit. Butit still exists two shortcomings, inefficient in incremental data extraction and poor adaptabilityin heterogeneous auditing environment. In addition, there are strong practical needs fornational audit, it aims to analyze audit data all of the provinces and municipalities, mineimplicit problem of related fields in a macro level, explore the impact factors in thedevelopment process of area, and provide a powerful and reliable data basis for relevantpolicies’ designing, promulgating and implement.To solve the above two points, this thesis proposes a new method of incremental dataextraction which is based on the division of tasks and deployed on multi-core servers. itequally distributes tasks which are waiting to different processes according to measurementby genetic algorithm and task-division algorithm, finally shortens the execution cycle ofwhole tasks. and can significantly enhances the extraction efficiency. And then, thesisanalyzes the requirements of national audit, for the current difficulties, researches dataintegration mechanism for the national audit which can quickly update the central datawarehouse by sending incremental combined with the existing network audit system on thisbasis, and make the data from all provinces be synchronous oriented to business logic bysynchronization module, to ensure that the data in warehouse of national audit office is fullysynchronous..Finally, the system for data integration oriented to national audit and identificationmethod based on task-division are checked by experiments, and the experimental results areanalyzed to verify the correctness and efficiency of the model, the cycle of method based ontask-division is less than existing identify methods, its speedup can reach49%, and achievethe design goals.