Research And Application Of Data Mining In Identifying Financial Fraud

Financial fraud has always been a headache for the financial sector and enterprises in various countries.Various countries have adopted various measures to prevent the occurrence of financial fraud.The traditional way to identify financial fraud is to audit the financial statements made by the CPA,and to express their opinions.This process depends largely on the professional judgment of the CPA.This approach has an unavoidable contradiction,the contradiction between the efficiency and effectiveness of audit identification.At present,the enterprise information is booming,and the transaction between enterprises has been realized through paperless and network.The traditional audit method is developed in the paper based office environment.In today's information environment,the efficiency and effectiveness of many methods have been greatly affected,the contradiction between efficiency and effectiveness of the audit is more prominent.Data mining has a unique advantage in identifying financial fraud.Compared with professional experience,using data mining techniques to identify financial fraud,to shorten the reaction time,the rational allocation of resources,reduce the subjective and arbitrary judgment and improve the accuracy of judgment has its unique advantages.This is for regulators to strengthen the crackdown,improve regulatory efficiency;investors reduce investment losses;auditors to control audit risk,reduce legal proceedings have important practical value.As an advanced analytical program for auditing,the data mining technology has the potential to solve the contradiction between effectiveness and efficiency when it is used as an assistant certified public accountant for fraud identification.This paper analyzes the existing data mining in financial fraud audit research,on this basis,put forward a set of application data mining to find the model of financial fraud.This paper selects the relevant information of internal management in the real life of listed companies published financial reports and enterprises for the raw data,carries on the pretreatment,convert it to the index system for comparison,which is then discretized using ChiMerge algorithm,which can be used for data mining using the data,finally achieve the financial fraud warning using decision tree algorithm,and according to the particularity of the fraud data on the decision tree algorithm is improved,making it more efficient.After the establishment of data mining to identify the financial fraud model,this article will use the programming language to achieve it,to convert it into practical use of the program.In this paper,the model is compared with the previous research,which proves the effectiveness of the C hiMerge algorithm and the improved decision tree algorithm.