Earning management estimation and prediction using machine learning: A systematic review of processing methods and synthesis for future research

Abstract

The present study highlights earning management optimization possibilities to constrain the events of earning management and financial fraud. Our study investigates the existing stock of knowledge and strand literature available on earning management and fraud detection. It aims to review systematically the methods and techniques used by prior research to determine earning management and fraud detection. The results indicate that prior research in earning management optimization is diverged among several techniques and none of these techniques has provided an ideal optimization for earning management. Further, the results reveal that earning management determinants are complex based on the type and size of business entities which complicate the optimization possibilities. The current research brings useful insights for predicting and optimization of earnings management and financial fraud. The present study has significant implications for policymakers, stock markets, auditors, investors, analysts, and professionals.