Knowledge-based Financial Statement Fraud Detection System: Based on an Ontology and a Decision Tree

Objective: Financial statements are of substantial significance to investors and other players in financial markets. Yet, these statements are sometimes misinforming and misleading. Thus, a potentially intricate issue for financial decision-makers is the prediction and detection of fraudulent financial statements. Method: This research investigates the relationship between earnings management and fraudulent financial statements in Firms, listed in Tehran Stock Exchange in order to help identify fraudulent financial statements in the time interval from 2009 to 2016. In other words, the objective of this study is to evaluate the ability of the popular discretionary accrual models to detect fraudulent financial statements. For this mission, 189 companies (21 fraudulent companies and 168 non-fraudulent) are selected as the research sample. The data mining methods employed in this research are Decision Trees (REPTree), Artificial Neural Networks (ANNs), and Bayesian Networks. We evaluate the ability of 7 measures derived from the extant discretionary accruals models to detect the existence of fraudulence. Results: The obtained results indicate that among all data mining methods, the Decision Trees method and among all accruals models, the Modified Jones model with book-to-market ratio have the greatest relationship with fraudulent financial statements.

This study aims to develop a better Financial Statement Fraud (FSF) detection model by utilizing data from publicly available financial statements of firms in the MENA region. We develop an FSF model using a powerful ensemble technique, the XGBoost (eXtreme Gradient Boosting) algorithm, that helps to identify fraud in a set of sample companies drawn from the Middle East and North Africa (MENA) region. The issue of class imbalance in the dataset is addressed by applying the Synthetic Minority Oversampling Technique (SMOTE) algorithm. We use different Machine Learning techniques in Python to predict FSF, and our empirical findings show that the XGBoost algorithm outperformed the other algorithms in this study, namely, Logistic Regression (LR), Decision Tree (DT), Support Vector Machine (SVM), AdaBoost, and Random Forest (RF). We then optimize the XGBoost algorithm to obtain the best result, with a final accuracy of 96.05% in the detection of FSF.

The study examines conceptually the detection of financial statements fraud using fraud pentagon. The study discusses the relationship between fraud pentagon and financial statement fraud. The methodology adopted in this study is library research whereby relevant and extant literature related to the elements in fraud pentagon were reviewed in relation to financial statements fraud. In the search for literature, it was observed that the various elements of fraud pentagon such as pressure (financial targets, financial stability, institutional ownership and external pressure), opportunity (ineffective monitoring and nature of industry), rationalization (auditor switching and change of director), competence (the changes in the company directorship and the proportion of independent directors in the board) and arrogance (frequent number of CEO's picture and CEO duality) significantly affect financial statements fraud. This study recommends that fraud pentagon should be given the statutory support in order to address financial statements fraud.
 
 Received: 27 August 2023 / Accepted: 19 October 2023 / Published: 5 November 2023

Auditing practices nowadays have to cope with an increasing number of fraudulent financial statements. Data mining techniques can facilitate auditors in accomplishing the task of detection of fraudulent financial statements (FFS). Considering the character of FFS, this paper designs a FFS detection model based on support vector machine. To perform the experiment, we choose 44 FFS according to the auditing reports and 44 non-fraudulent financial statements(non-FFS) according to some specific standards from listed companies in China during 1999–2002 as training data set. Similarly, 73 FFS and 99 non-FFS during 2003–2006 are chosen as testing data set. We train the model using training data set and apply the trained model to the testing data set, good experimental results are obtained.

Detecting evolutionary financial statement fraud

This research aims to examine the fraud hexagon theory in the detection of fraudulent financial reporting that occurs in a company. According to the hexagon fraud theory, fraud occurs due to 6 factors, namely pressure/stimulus, opportunity, capability, ego/arrogance, rationalization, and collusion. These 6 factors are proxied with 10 proxies, namely: financial stability, financial targets, external pressure, change in director, change in auditor, ineffective monitoring, industry nature, number of CEO photos, political connections, and collusion. The population in this research are companies listed in the LQ 45 index period I February-July 2017-2021. The sample in this study is 88 samples. The data analysis approach used is statistic descriptive, classic assumption test, and multiple regression analysis using SPSS 25. The results of this study reveal that the nature of industry and external pressures have a negative effect on the detection of financial statement fraud, and change in auditor positively effect on the detection of financial statements fraud. While the variables of financial stability, financial targets, political connections, change in directors, number of CEO photos, ineffective monitoring, and collusion have no effect in detecting financial statements fraud. This result can be used for companies management as their references about what makes fraudulent financial statements occurs, so they can make preventive action for their companies

This study aims to examine the effect of Fraud Hexagon on Financial Statement Fraud in mining companies listed on the Indonesia Stock Exchange (IDX). Fraud Hexagon consists of pressure, opportunity, rationalization, capability, arrogance, and collusion, while Financial Statement Fraud is measured using the Beneish M-Score. This study uses a quantitative approach with logistic regression analysis and purposive sampling techniques based on the companies' financial statements and annual reports for the observation period. The results show that simultaneously, the Fraud Hexagon influences Financial Statement Fraud, but partially, not all variables have a significant influence. These findings are expected to contribute to the detection and prevention of financial statement fraud in the mining secto. Keyword: Fraud Hexagon, Financial Statement Fraud, Beneish M-Score, Pertambangan.

The purpose of this study is to respond to research gaps and inconsistencies in research findings by empirically testing how financial targets and external pressures affect the detection of financial statement fraud. The study utilizes a descriptive research design with a quantitative approach. A method of purposive sampling was used, focusing on companies listed on the Jakarta Islamic Index (JII) from 2019 to 2023. Statistical analysis was done using multiple regression analysis. The results show that financial targets have a significant effect on the detection of financial statement fraud, while external pressures do not have a significant effect. However, both financial targets and external pressures have a simultaneous effect on the detection of financial statement fraud. The practical implications, financial targets, and external pressures can have an impact on the detection of fraud within the organization. Under these conditions, fraud may be more difficult to detect because the organization's or individual's focus is more on achieving targets or meeting external expectations. Therefore, companies need to have a strong internal control system and independent audits to help identify and manage potential risks of financial statement fraud.

Financial statement fraud has been one of the biggest challenges in the modern business world. Financial accounting fraud detection (FAFD) has become an emerging topic of great importance for academic, research and industries. In this paper, the effectiveness of Data Mining (DM) classification techniques in detecting firms that issue fraudulent financial statements (FFS) and deals with the identification of factors associated to FFS are explored. Our study investigates the usefulness of Data Mining techniques including Decision Trees, Neural Networks and Bayesian Belief Networks in the identification of fraudulent financial statements. At last, we compare the three models in terms of their performances.

The selection of quality rice seeds is vital for maximizing agricultural productivity and sustainability. This study develops an expert system for rice seed selection based on intelligent data processing using the Knowledge-Based System (KBS) and Decision Tree methods. KBS encodes expert knowledge to evaluate seed quality and environmental compatibility, while Decision Tree algorithms classify and predict optimal seed choices. Experimental results demonstrate the system's accuracy in recommending suitable seeds, reducing selection time and effort. This research highlights the potential of artificial intelligence in enhancing decision-making processes in modern agriculture.

With the prevalence of Internet AI technology, financial fraud becomes an imperative problem, especially in the context of machine learning. Technologies such as deep learning and natural language processing provide effective tools for detecting fraud with the guidance of financial statements, improving the efficiency and accuracy of data analysis, and helping to ensure financial safety. In this study, we propose a sophisticated representation learning method to detect financial statement fraud by tracking the detailed changes in the company’s Management Discussion and Analysis (MD&A) documents over time. Unlike traditional word frequency methods, we align paragraphs between consecutive disclosures based on their similarity at the paragraph level and categorize them into three types: added, deleted, and matched. Next, we create multivariate change trajectory representations based on fraud-related word categories. Finally, we use these word-level change trajectories to design a fraud detection model and compare it with several traditional models as well as the latest Time-Series Foundation Models. Experiments on 24 years of financial report data, from 1995 to 2019, show that our representation learning method significantly improves the performance of financial statement fraud detection across 11 different machine learning models, consistently outperforming traditional word frequency methods. Our method opens a new paradigm for feature engineering in financial statement fraud detection. Our code can be found at https://github.com/LittelStudent/Financial-Statement-Fraud-Detection-ParaEmb-FraudW2V .

This study aims to analyze the influence of the fraud Pentagon elements on detecting financial statement fraud in BUMN listed on the Indonesia Stock Exchange (IDX) for 2020-2022. The study population was all state-owned enterprises BUMN listed on the Indonesia Stock Exchange. The sample was selected using a purposive sampling method, resulting in 18 companies. Data analysis used multiple regression analysis. The results showed that financial stability and personal financial needs were proven to have no significant effect on financial statement fraud. Financial targets, changes in director and auditor, and the number of CEO photos in the annual report showed a significant positive effect on financial statement fraud. Meanwhile, ineffective monitoring and dualism positions showed a significant negative effect. These findings indicate that several elements in the fraud pentagon framework significantly affect financial statement fraud in BUMN in Indonesia, with the model explaining 62.6% of the variance in fraud detection. ABSTRAK Penelitian ini bertujuan menganalisis pengaruh elemen fraud pentagon terhadap pendeteksian kecurangan laporan keuangan pada BUMN yang terdaftar di Bursa Efek Indonesia (BEI) periode 2020-2022. Populasi penelitian adalah seluruh badan usaha milik negara BUMN yang terdaftar pada Bursa Efek Indonesia. Sampel dipilih menggunakan purposive sampling method sehingga sejumlah 18 perusahaan. Analisis data menggunakan analisis regresi berganda. Hasil penelitian menunjukkan financial stability dan personal financial needs terbukti tidak berpengaruh signifikan terhadap kecurangan laporan keuangan. Financial targets, change in director, change in auditor, dan jumlah foto CEO dalam laporan tahunan menunjukkan pengaruh positif signifikan terhadap kecurangan laporan keuangan. Sementara itu, ineffective monitoring dan dualism position menunjukkan pengaruh negatif signifikan. Temuan ini mengindikasikan bahwa beberapa elemen dalam kerangka fraud pentagon secara signifikan mempengaruhi kecurangan laporan keuangan pada BUMN di Indonesia, dengan model menjelaskan 62,6% varians dalam pendeteksian fraud.

The objective of this research is to analysis the effect of auditor characteristics that represented by professional scepticism and auditor experiences toward financial statement fraud detection through artificial intelligence in data analytic. This study builds upon previous research where respondents were auditors who had used big data analytics and incorporate artificial intelligence, specifically big data analytic as mediator between auditor characteristics and the detection of financial statement fraud. Quantitative research was used in this research by obtaining primary data through 66 respondents from auditor at several public accounting firms in the Jakarta area. Our findings reveal that auditor experience has direct influence effect on financial statement fraud detection while professional scepticism has no effect. But surprisingly, the role of artificial intelligence in data analytics is only able to mediate professional scepticism on financial statement fraud detection, while this does not apply to auditor experience. This study provides empirical evidence that the role of artificial intelligence, namely big data analytic, can act as a catalyst that transforms professional scepticism from mere doubt to evidence-based action, by enhancing the auditor’s ability to identify data patterns and anomalies, reducing cognitive bias and increasing objectivity.

Financial Feature Embedding with Knowledge Representation Learning for Financial Statement Fraud Detection

The rise of financial fraud, not only in India but also around the world, is a major problem. Credit card transactions have been steadily increasing in recent years, as have internet payments like Unified Payments Interface and debit card transactions, which have been increasing every day. As a result, fraud is increasing, and it has become easier for fraudsters to commit fraud. Recent research indicates the practicality of machine learning algorithms for recognizing payment fraud. Since credit cards are a common target, online fraudsters simply commit fraud because of excess use of e-commerce and other platforms, which has led to an increase in online payment methods, resulting in a greater risk of online fraud. Many researchers have started using machine learning algorithms to detect fraud. The primary objective of this study is to create a better fraud detection algorithm by analyzing payment transactional data, focusing on customers' transactional data and their payment behavioral patterns. The study proposes a framework that groups cardholders based on the volume of their transactions. Financial fraud is the biggest threat to world economies in terms of financial losses each year. The task of accurately and efficiently detecting fraud is, therefore, very challenging due to the high volume of transaction data and its complex nature. This paper presents a hybrid machine learning approach using logistic regression, decision trees, and random forests that could help enhance the accuracy and reliability of fraud detection systems. The study used an available credit card fraud dataset, with a focus on feature engineering and model evaluation to compare individual algorithms versus their ensemble. Experimental results show that the ensemble outperformed the individual classifiers in terms of precision, recall, and overall accuracy. This paper underlines the potential of hybrid machine learning techniques in improving fraud detection systems and proposes avenues for further research.