Research on RPART AdaBoost Model for Corporate Financial Reporting Based on Data Mining
In order to improve the authenticity of accounting information,maintain a good order in the securities market,and expand the application of data mining techniques in identifying the characteristics of financial fraud models,using 36 manufacturing industry enterprises with financial reporting violations as the research object and selecting companies of the same size and quantity as control samples,an RPART AdaBoost model is constructed to analyze the improvement effect of financial and non-financial indicators on the accuracy of identifying financial reporting violations in the model.The results indicate that financial indicators are the main variable for identifying violations,and non-financial indicators have not passed the Man Whitney U-test.The RPART AdaBoost model has a recognition accuracy of 90.32％for all samples.
NETL
NSTL
万方数据
