Adaptive Identification and Warning Method for Financial Fraud Behaviour based on Deep Q-Learning

Purpose This paper aims to integrate the latent semantic features of annual report text with accounting indicators to construct a financial fraud identification model, and quantitatively analyze the impact of different corporate risks on financial fraud behavior in different industries, providing a reference for identifying financial fraud. Design/methodology/approach This paper obtains 3,860 corporate annual report samples and accounting indicators from 2001 to 2020 through crawlers and the CSMAR database as our experimental subjects. By integrating latent semantic features with accounting indicators and textual language features, a new indicator system group is constructed. Based on this indicator system group, multiple model identification effects are compared and a stacking-based enterprise financial fraud identification model is constructed. In addition, an econometric model is established to verify the impact of latent semantic features related to enterprises on corporate financial fraud. Findings The experimental results show that the constructed stacking-based enterprise financial fraud identification model performs better than other machine learning models and can effectively identify financial fraud. The econometric model established for the latent semantic information of annual reports explains the impact of different corporate trends on fraud behavior in different industries. Originality/value This paper combines the textual latent semantic features of annual reports with accounting indicators, expands the scope of data analysis, introduces the idea of ensemble learning, updates the financial fraud identification algorithm and constructs an econometric model for further analysis, providing a reference for financial fraud identification.

This study investigates the applicability of the Fraud Hexagon Theory (S.C.C.O.R.E. Theory) in explaining financial fraud within Chinese accounting firms, addressing a critical gap in understanding how Stimulus, Capability, Collusion, Opportunity, Rationalization, and Ego (S.C.C.O.R.E.) influence fraudulent behavior in this context. Employing a quantitative approach, we collected data from 238 professionals in Chinese accounting firms via structured questionnaires, ensuring ethical compliance through anonymized responses and informed consent. Structural Equation Modeling (SEM) was utilized to analyze the relationships between the six fraud determinants and their interaction with Corporate Governance as a moderating variable. Key findings reveal that Stimulus, Capability, Collusion, Opportunity, and Ego exhibit significant positive effects on Fraud Behavior, while Rationalization showed no statistical significance. Notably, Corporate Governance negatively moderates the relationship between Opportunity and Fraud Behavior, suggesting that robust governance mechanisms can mitigate exploitable opportunities. This study contributes empirically by validating the Fraud Hexagon Theory in China’s unique institutional and cultural setting, contrasting with prior Western-centric studies, and offers actionable insights for policymakers and accounting firms to refine anti-fraud frameworks. Practically, the results underscore the need for ethics training targeting Ego-driven rationalizations and regulatory reforms to strengthen governance oversight. Methodologically, the research advances fraud literature by integrating SEM with a context-specific moderating analysis, addressing limitations in traditional linear models. Ethical considerations, including participant confidentiality and voluntary participation, were rigorously upheld throughout the study. These findings provide theoretical and practical guidance for corporate managers, regulators, and educators in combating financial fraud and fostering professional integrity in China’s rapidly evolving financial sector.

Artificial Intelligence in Financial Fraud Detection: A Comprehensive Approach to Enhancing Financial Security The rise of artificial intelligence (AI) offers an opportunity to significantly strengthen financial security by combating financial fraud, which has become increasingly complex and widespread. Traditional detection methods are often insufficient in identifying and preventing fraudulent activities, prompting a shift towards AI-based solutions. This study explores the application of AI, particularly machine learning algorithms, in improving the accuracy and efficiency of fraud detection. By analyzing large financial datasets, AI can detect anomalies that may indicate fraudulent behavior more effectively than traditional approaches. This research adopts a two-phase methodology. The first phase involves a thorough review of existing financial fraud detection methods, comparing traditional techniques with AI-based models to identify gaps. Various machine learning approaches, including supervised, unsupervised, and deep learning algorithms, are reviewed for their effectiveness in detecting fraud. The second phase involves developing and testing an AI model to identify fraudulent patterns within transactional data. The model uses machine learning algorithms to process vast datasets and detect deviations from typical financial behaviors, flagging potentially fraudulent activities. The expected results indicate that AI systems can outperform traditional fraud detection methods by significantly reducing false positives and improving the detection rate of genuine fraud. This reduction in false positives is vital for financial institutions, as it reduces unnecessary investigations and saves valuable resources. Additionally, enhanced fraud detection protects both institutions and consumers from financial losses.The findings of this study aim to provide financial institutions with practical insights into the implementation of AI-driven fraud detection systems. Furthermore, the research highlights the need for continuous refinement of AI models to adapt to the evolving nature of financial fraud. By leveraging AI technologies, financial institutions can revolutionize their approach to fraud detection, making financial systems more secure and responsive to emerging threats.

Objective –The specific objectives to be achieved in this research are to (1) test the Beneish model (1999) as a predictor of reliable manipulation or indicator to reveal signals or symptoms of financial reporting fraud. (2) Test whether corporate governance effectiveness reduces financial statement fraud.Design/Methodology –The research analyzed publicly listed non-financial firms listed in Indonesia's stock exchange (IDX) from 2017 to 2022. Simple random sampling was used to select 86 companies per year, resulting in 617 year-firm observations. Data analysis was carried out by logistic regression multiple with panel data. The test was carried out by adopting the Beneish benchmark index (Z-score) model which is effective for predicting indications of fraudulent financial statements. Results –The test results show that there is 99.8% of incidents indicated financial statement fraud and the other 0.2% did not indicate financial statement fraud. Corporate governance mechanisms consisting of institutional ownership, audit committees, and audit quality does not mitigate fraud. It is anticipated that the research's contribution will serve as a practical guide and reference for the examination of indicators of fraud in corporate finances. The findings of this study are expected to provide valuable insights for policymakers and regulators seeking to improve corporate governance practices.Research limitations/implications –This research does not incorporate the competency and capability specifications of governance structures such as the board of commissioners, audit committees, and external auditors. Future investigations are anticipated to delve more profoundly into the internal and external corporate governance capabilities and specifications.Novelty/Originality –This study reveals that, on the whole, a significant proportion of companies are implicated in fraudulent activities, while corporate governance mechanisms are deemed inadequate in overseeing management activities conducive to fraud. The outcomes of this research challenge the prevailing paradigm of corporate governance as an effective deterrent against fraudulent behavior.

In the current era of technology, the world at large has been benefitted from the technological progression in almost all the sectors. The advancement of technology, on the one hand, has brought about many boons to society at large, but on the other hand, the fraudsters have found new and faster ways to commit the crime. Hence there is a dire need to introduce new methods to detect and control frauds in the current scenario oftechnological development. The sector of society that is most affected by these Information Technology frauds is the financial sector. These financial frauds can range from small threats like non-payment for e-commerce orders to major coercions like the exposure of customer’s credit or debit card details. When fraud is detected, the organizations end up paying for the heavy losses, and their reputation decays. Hence the fraud detection and prevention methods need to adopt new-age technologies like artificial intelligence (AI) and machine learning (ML) to stop the fraudsters from committing heinous financial crimes. Once fraud is committed, one can only take corrective actions which involve tracking down the culprit and try to recover the defrauded amount. However, the much better option is to prevent the fraud before it occurs, and this exactly is where the role of AI steps in, which is supported by ML, comes into the picture and provides the best solutions available so far. ML can identify legitimate (genuine) and fraudulent behaviors by comparing themwith previously unseen fraud tactics. As far as the role of AI is concerned concerning the use of the models defined by ML, AI uses two models to identify fraudulent practices. One is a supervised model, while the other one is an unsupervised model. Though even after employing ML and AI, certain challenges complicate the fraud detection process, such as changing of fraud patterns by fraudsters, lack of technical knowledge to customers, model interpretations, and time needed to develop and review the already set models. However, ML models and AI increase system efficiency and reduce the instances of fraud occurrence.

In recent years, with financial fraud events coming one after another, the legitimate rights and interests of consumers have been seriously damaged. To solve the problem that financial fraud behaviors are more difficult to identify and more hidden, the study proposes a financial fraud identification model based on stacked integrated learning model and accounting metrics. The model adopts a multi-level feature extraction method and combines multiple base learners and logistic regression to identify financial fraud based on accounting metrics. In this study, the accuracy of the Stacking model is 85.23%, which is 5.32% higher than that of the other 5 Stacking models. Compared with other feature models, the accuracy and F1-score of the recognition model based on accounting index are increased by 18.68% and 19.11% on average. The study indicates that the financial fraud identification model based on Stacking as well as accounting metrics has good performance along with low under reporting rate for fraudulent firms. The study explores financial fraud from multiple dimensions based on annual reports, which can effectively reduce the human economic losses caused by financial fraud and assist the development of anti-fraud behavior.

Fraud and financial crime continue to pose significant challenges for private companies, making prevention efforts essential through the identification of their root causes. This study examines the relationship between situational crime prevention and financial fraud, with a focus on employee tenure as a moderating variable. An empirical research method was employed, with data collected from employees of PT Budi Jaya and its subsidiaries through questionnaires. A total of 60 valid responses were analyzed using WarpPLS. The findings reveal that situational crime prevention—measured through indicators such as minimal benefits, high risks, limited opportunities, and strict sanctions—has a significant negative impact on financial fraud. Additionally, the study demonstrates that employee tenure moderates the relationship between situational crime prevention measures and financial fraud. This research makes a theoretical contribution to situational crime prevention theory by incorporating the concept of rationalization, which fraud perpetrators often use to justify their actions. The study offers practical insights, suggesting that companies and regulators should implement robust internal control systems to minimize opportunities for fraud and establish clear, stringent rules and sanctions to deter fraudulent behavior.

In recent years, there are often high premium acquisitions based on high-performance commitment in corporate mergers and acquisitions. The performance commitment is intended to alleviate the information asymmetry between the two parties of mergers and acquisitions, but the pressure brought by the high-performance commitment often triggers the company’s financial fraud behaviour. This paper will use the case study method to select the classic case of Aerospace Communications merger and acquisition of the HiPAD, in-depth analysis of the merged party HiPAD of the financial counterfeiting means, and corresponding to the identification of signs of financial fraud research. This paper analyses the abnormal signs in the public disclosure information of HiPAD from the unique perspectives of financial information identification and non-financial information identification, parent company and subsidiaries, providing a reference direction for the financial fraud identification path. Finally, it puts forward a universal proposal for the prevention of corporate financial fraud in the context of the pressure of the performance commitment, which provides a feasible method for the similar mergers and acquisitions and high-performance commitment cases.

In recent years, with the development of the China’s economy, in order to seek business growth points, the number of mergers and acquisitions of enterprises has increased rapidly. In the process of mergers and acquisitions, there are often some financial fraud behaviors, such as fictitious income and false statements, to push up the merger and acquisition premium, which will greatly damage the stability of the financial market and the legitimate rights and interests of investors. This paper adopts the financial fraud case of Changyuan Group's acquisition of Shanghai Heying Electromechanical Technology Co., LTD. in 2016 as an example, and combines the fraud triangle theory to analyze the motivation of financial fraud. According to the punishment announcement and other information, this paper analyzes the identification of financial fraud information with a three-dimensional model, analyzes and presents the consequences of financial fraud, and proposes the suggestions and enlightenment for subsequent M&A cases.

This study investigates the effect of the New Fraud Diamond components, financial target, nature of industry, earnings growth, and change of directors on financial statement fraud among construction companies listed on the Indonesia Stock Exchange (IDX) during the period 2020–2023. Guided by the New Fraud Diamond framework and Agency Theory, the research employs a quantitative approach using logistic regression analysis on 80 firm-year observations. The results show that all four independent variables have a significant effect on the likelihood of financial statement fraud. Specifically, low profitability (financial target), low receivables turnover (nature of industry), high earnings growth, and the presence of director changes are all positively associated with fraudulent financial reporting. These findings confirm the applicability of the New Fraud Diamond model in emerging markets and highlight the importance of identifying fraud risk factors for early detection. The study contributes to the literature by offering empirical evidence on the role of pressure, opportunity, integrity, and capability in shaping corporate fraud behavior within the Indonesian construction sector.

Financial fraud detection has been an urgent technical demand in cyberspace. It highly relies on clear extraction and deep representation toward complex relationships inside financial social networks. As consequence, this study combines both knowledge graph and deep learning to deal with such issue. Thus, an intelligent financial fraud detection model based on knowledge graph guidance and deep neural network is proposed in this paper. First, a new knowledge graph based on financially related systems is constructed, which includes multiple entities and relationships. Then, an adversarial learning-based neural network structure is formulated to extract financial attributes. Finally, the detection results can be output according to the extracted factors. Empirically, the proposal is implemented on a real-world dataset for performance evaluation. The experimental results show that it has higher accuracy and effectiveness compared to traditional fraud detection methods. The proposed detection model can not only identify known fraudulent behaviors, but also predict potential fraud patterns based on consumer habits, thereby improving the security and reliability of financial transactions. It can also update the knowledge graph in real-time, enabling it to cope with emerging fraud methods and variants.

This article focuses on fraudulent behaviour and patterns as well as ways of detecting such patterns by using Big Data. The study analyses scientific articles to examine types of financial fraud and their detection techniques as well as develops a model that is based on factors characterizing fraudulent credit card transactions made across USA. Regression analysis, correlation and descriptive statistics analysis is applied. Statistically significant results are found indicating a causal relationship between fraudulent transactions and transactions made in Alaska, during the month of October and on a Thursday. Although, the impact of these relationships is relatively small. Expanding the dataset with more numerical variables that could be used for identifying fraudulent transactions is advised for future research as to better the overall fit of the model.

Financial fraud has become increasingly sophisticated as corporations expand digital operations, adopt blockchain-based systems, and intensify ESG reporting practices. Traditional forensic accounting techniques—reliant on manual audits and ratio-based analysis—are often unable to detect complex, multimodal fraud patterns. This study develops an AI-driven hybrid forensic accounting framework that integrates financial statement ratios, blockchain transaction anomalies, and ESG governance and sentiment indicators to enhance fraud detection. Using a simulated but realistic multimodal dataset of 2,000 firms, five supervised machine learning models are evaluated, with XGBoost achieving the highest predictive performance (F1-score = 0.80; ROC-AUC = 0.92). Comparative analyses reveal that blockchain-derived features improve detection accuracy by approximately eight percentage points, while ESG governance and sentiment indicators emerge as significant predictors of fraudulent behavior. The hybrid model significantly outperforms financial-only, blockchain-only, and ESG-only models, confirming that fraud is inherently multidimensional. Explainable AI (SHAP) further ensures interpretability without degrading model performance, strengthening transparency and forensic defensibility. The findings demonstrate that an integrated AI–Blockchain–ESG approach provides a robust, transparent, and scalable solution for modern fraud detection challenges and offers a theoretical and practical foundation for next-generation forensic accounting systems.

This study aims to analyze the application of the Fraud Diamond Theory in detecting fraudulent financial statements in manufacturing companies in Indonesia. The Fraud Diamond Theory, which consists of four key elements—pressure, opportunity, rationalization, and capability—provides a comprehensive framework for understanding the factors that drive individuals to commit financial fraud. Using a sample of financial data from selected manufacturing firms, this research examines how these four elements contribute to fraudulent behavior and evaluates their effectiveness in identifying financial statement irregularities. The study utilizes quantitative methods, employing financial ratio analysis, statistical tests, and data on corporate governance to assess the likelihood of fraud within these companies. The results indicate that the elements of pressure and opportunity play a significant role in facilitating fraudulent activities, while rationalization and capability also contribute but are more difficult to quantify. The findings suggest that incorporating Fraud Diamond indicators into corporate governance and internal audit systems can enhance the early detection of financial fraud in manufacturing firms in Indonesia. This research contributes to the growing body of literature on fraud detection, offering insights for regulators, auditors, and corporate stakeholders to strengthen financial oversight and prevent financial misstatements.

We develop and test a theoretical model to investigate the effects of faultlines within the top management team (TMT) on corporate financial fraud. We propose that TMT faultlines can generate mutual monitoring among factional subgroups in the executive suite, which reduces fraudulent behavior. We also examine the contingent roles of subgroup configuration and the TMT members’ tenure overlap in shaping the relationship between TMT faultlines and financial fraud. The mutual monitoring effect is likely to be stronger when the TMT has a balanced subgroup configuration and shorter TMT members’ tenure overlap. We test our argument in the context of publicly listed firms in China. This article extends the mutual monitoring perspective of corporate governance and has important research implications for the corporate financial fraud literature.