Systematic Literature Review (SLR): Analysis of Financial Economic Reporting Fraud Detection Models Based on Hybrid Machine Learning and Psychological Factors

Financial reporting fraud is a major challenge that threatens the integrity of financial markets. Hybrid Machine Learning (HML) offers great potential in detecting increasingly complex fraud, but its integration with psychological analysis is still limited. This study uses the Systematic Literature Review (SLR) approach with the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) method to identify trends, challenges, and opportunities in the application of HML for detecting financial reporting fraud. Data were collected from various leading academic databases, such as ScienceDirect, Web of Science, IEEE Xplore, SINTA, SCOPUS, and ProQuest, with relevant keywords. The selection process was carried out through the stages of identification, screening, eligibility evaluation, and inclusion, resulting in 27 main articles published between 2017-2025 from various countries. This study found that financial reporting fraud detection has developed significantly with the integration of HML and psychological factor analysis. Most studies focus on quantitative approaches based on Machine Learning (ML), Deep Learning (DL), and Big Data Analytics , with the main variables being financial ratios, corporate governance, and psychological factors. However, a multidisciplinary approach that combines AI techniques, forensic auditing, and psychological insights is still needed. These findings contribute to identifying research gaps and directions for the development of more comprehensive fraud detection models.