Predicting Accounting Fraud in Publicly Traded Chinese Firms via A PCA-RF Method
Donger Chen · Advances in computer science research · 2022
Financial fraud occurs from time to time and gradually becomes a worldwide problem with the expansion of the international capital markets and the rise of the information industry economy under the Internet eco-system.This paper provides a methodology for predicting financial fraud using basic financial data.The methodology is based on PCA-RF.Different from traditional methods such as logistic regression and support vector machine, we creatively proposed a PCA-RF model, first using principal component analysis to reduce the dimensionality of the data, then using grid search to optimize the random forest model, and finally directly selecting the raw financial data from the financial statements for direct analysis.We compare the analysis results with random forest and neural network methods, and the study finds that the PCA-RF model is superior for predicting domestic financial fraud in China.In this paper, we use an ensemble learning approach to introduce the PCA-RF method into the field of prediction of financial fraud for listed companies.