Research And Practice Of Financial Big Data Audit Technology

With the continuous development of computer technology,big data technology is becoming more and more mature and perfect.The wide application of big data technology has brought profound influence to social operation and development.The information technology such as big data and Internet plus promotes the rapid development of the modern economy,but also promotes the continuous improvement of the national governance mechanism and the improvement of governance level,gradually promoting the modernization of national governance.Audit institutions as an important part of the national governance system,of course,the transformation of audit supervision mode should also keep pace with the times.Making good use of big data technology to carry out audit work is an inevitable requirement for the state to promote data transparency,sharing and opening,and enhance the national data capacity.Audit is an important guarantee to promote the modernization of national governance.Traditional data audit technology can not meet the requirements of fast,accurate and low cost in today's society.Big data audit is the only way to achieve full coverage of national audit.On the basis of fully investigating the knowledge of financial data audit and carefully analyzing the relevant needs,this paper constructs a financial big data audit platform through offline and online data acquisition methods.The platform uses big data technology to achieve a comprehensive audit including the overall budget,department budget and special budget.It also makes a deep analysis of financial data,finds the internal relationship between data and excavates a new audit model,combining with relevant algorithms such as association rules and regression prediction in data analysis.The main work of this paper includes:(1)After analyzing the principle of Apriori algorithm in detail,aiming at the generation process of candidate itemsets,a method of adding filter sets is proposed.By using known infrequent itemsets,the number of candidate itemsets can be further reduced,and even the number of scans of data sets can be reduced,so as to improving the performance of the algorithm.Through association rules algorithm,frequent items and association rules are found in a large number of data,and financial data audit model is established according to the useful rules mined.(2)When Gauss kernel function is used to give higher weights to adjacent points for the local weighted linear regression algorithm,it is necessary to set a parameter K to express the change speed of weights.In order to weigh the training error and test error,a calculate method of parameter k is proposed,which involves the calculation of the correlation between the real value and the predicted value and the sum of squares of errors.This paper studies and analyses another gradient boosting tree GBDT algorithm for regression problems,and compares with linear regression.By using the forecast model which is established by the local weighted linear regression algorithm,the next year's revenue budget is predicted according to the financial data of revenue budget and execution in previous years.It provides reference and data support for decision makers to make the next year's plan correctly;(3)The design and implementation of financial budget auditing system includeing the functions of routine data query,report query,comparison and error correction.In addition,it achieves comprehensive Auditing from the general budget,department budget and special budget.(4)The function test of the financial budget auditing system has been carried out,and the result indicates basically that the system meets the construction targets and requirements.