Research On Financial Fraud Recognition Of Listed Companies Based On Data Mining Technology

The financial fraud of listed companies undermines the open and transparent information of securities market information,this will mislead the report of the user because of false information and make them can not do the right judgments and decision-making,and then resulting in losses from investors and other stakeholders,affecting investors' confidence in securities investment.Because of this,the recognition of financial fraud has an important role.However,with the rapid development of China's securities market,the number of listed companies is also rising rapidly,such as the number of A-share listed companies has more than doubled in the decade from 2005 to 2014,this is a great challenge to the financial fraud recognition of listed companies.As a result of manual identification of the audit requirements of workers have higher professional quality,and this method also need to spend a lot of manpower and resources,so it is important and practical to find a new way to identify the financial fraud behavior of listed companies.This article applies data mining technology to the recognition of financial fraud in listed companies,based on the deep understanding of the theory of financial fraud,comprehensive application data mining technology to classify fraud and non-fraud companies.The article first introduces the related concepts of financial fraud and data mining through literature research,and a brief introduction to the commonly used tools for data mining.Secondly,use Boruta and Relief two feature selection algorithm respectively from 2007 to 2014 between the corrupt companies and non-fraud companies to filter.And then we use the logic regression,support vector machine,neural network and random forest four kinds of classification methods and the feature selection algorithm selected by combining the indicators to establish the model,and the four classification models were evaluated by model evaluation method.Finally,the article has made some suggestions to curb the financial fraud of listed companies.