Traditional anomaly detection methods often assume that data points are independent or exhibit regularly structured relationships,as in Euclidean data such as time series or image grids.However,real-world data frequen...Traditional anomaly detection methods often assume that data points are independent or exhibit regularly structured relationships,as in Euclidean data such as time series or image grids.However,real-world data frequently involve irregular,interconnected structures,requiring a shift toward non-Euclidean approaches.This study introduces a novel anomaly detection framework designed to handle non-Euclidean data by modeling transactions as graph signals.By leveraging graph convolution filters,we extract meaningful connection strengths that capture relational dependencies often overlooked in traditional methods.Utilizing the Graph Convolutional Networks(GCN)framework,we integrate graph-based embeddings with conventional anomaly detection models,enhancing performance through relational insights.Ourmethod is validated on European credit card transaction data,demonstrating its effectiveness in detecting fraudulent transactions,particularly thosewith subtle patterns that evade traditional,amountbased detection techniques.The results highlight the advantages of incorporating temporal and structural dependencies into fraud detection,showcasing the robustness and applicability of our approach in complex,real-world scenarios.展开更多
This study investigates the dark side of the non-fungible token(NFT)marketplace,with a focus on understanding the risks,and underlying factors driving fraud in the NFT ecosystem.Using the fraud triangle framework,this...This study investigates the dark side of the non-fungible token(NFT)marketplace,with a focus on understanding the risks,and underlying factors driving fraud in the NFT ecosystem.Using the fraud triangle framework,this study examines pressure,opportunity,and rationalization from individual and organizational perspectives.The research provides a comprehensive understanding of the contributing factors to NFT marketplace fraud by analyzing the reasons behind fraudulent actions.A conceptual framework is developed that includes ten propositions to aid in understanding the complexity of this issue.This study’s outcomes will assist policymakers in crafting efficient approaches to mitigate fraud within the NFT marketplace.展开更多
This research aims to understand the causes of fraud through the approach of the Fraud Triangle Theory,which includes three main factors:pressure,opportunity,rationalization.By using the Systematic Literature Review m...This research aims to understand the causes of fraud through the approach of the Fraud Triangle Theory,which includes three main factors:pressure,opportunity,rationalization.By using the Systematic Literature Review method from various relevant international journals,it analyzed systematically to identify patterns,trends,and theoretical contributions to efforts in detecting and preventing fraud.The results of this study show that the three factors in the Fraud Triangle Theory significantly contribute to the occurrence of fraud.Opportunity as the most dominant factor is caused by weak internal control systems and lack of oversight.In addition,economic pressure,and a permissive organizational environment,as well as the rationalization processes by individuals,also increase the tendency for a person to commit fraud.These findings emphasize the need for a comprehensive approach in fraud prevention strategies,through ethical values,the establishment of an organizational culture with integrity,and the implementation of more effective internal control and oversight.展开更多
Reducing the risk of fraud in credit card transactions is crucial for the competitiveness of companies,especially in Latin American countries.This study aims to establish measures for preventing and detecting fraud in...Reducing the risk of fraud in credit card transactions is crucial for the competitiveness of companies,especially in Latin American countries.This study aims to establish measures for preventing and detecting fraud in the use of credit cards in shops through analytical methods(data mining,machine learning and artificial intelligence).To achieve this objective,the study employs a predictive methodology using descriptive and exploratory statistics and frequency,frequency&monetary(RFM)classification techniques,differentiating between SMEs and large businesses via cluster analysis and supervised models.A dataset of 221,292 card records from a Latin American merchant payment gateway for the year 2022 is used.For fraud alerts,the classification model has been selected for small and medium–sized merchants,and the multilayer perceptron(MLP)neural network has been selected for large merchants.Random forest or Gini decision tree models have been selected as backup models for retraining.For the detection of punctual fraud patterns,the K-means and partitioning around medoids(PAM)models have been selected,depending on the type of trade.The results revealed that the application of the identified models would have prevented between 48 and 85%of fraud transactions,depending on the trade size.Despite the promising results,continuous updating is recommended,as fraudsters frequently implement new fraud techniques.展开更多
This paper proposes a novel hybrid fraud detection framework that integrates multi-stage feature selection,unsupervised clustering,and ensemble learning to improve classification performance in financial transaction m...This paper proposes a novel hybrid fraud detection framework that integrates multi-stage feature selection,unsupervised clustering,and ensemble learning to improve classification performance in financial transaction monitoring systems.The framework is structured into three core layers:(1)feature selection using Recursive Feature Elimination(RFE),Principal Component Analysis(PCA),and Mutual Information(MI)to reduce dimensionality and enhance input relevance;(2)anomaly detection through unsupervised clustering using K-Means,Density-Based Spatial Clustering(DBSCAN),and Hierarchical Clustering to flag suspicious patterns in unlabeled data;and(3)final classification using a voting-based hybrid ensemble of Support Vector Machine(SVM),Random Forest(RF),and Gradient Boosting Classifier(GBC).The experimental evaluation is conducted on a synthetically generated dataset comprising one million financial transactions,with 5% labelled as fraudulent,simulating realistic fraud rates and behavioural features,including transaction time,origin,amount,and geo-location.The proposed model demonstrated a significant improvement over baseline classifiers,achieving an accuracy of 99%,a precision of 99%,a recall of 97%,and an F1-score of 99%.Compared to individual models,it yielded a 9% gain in overall detection accuracy.It reduced the false positive rate to below 3.5%,thereby minimising the operational costs associated with manually reviewing false alerts.The model’s interpretability is enhanced by the integration of Shapley Additive Explanations(SHAP)values for feature importance,supporting transparency and regulatory auditability.These results affirm the practical relevance of the proposed system for deployment in real-time fraud detection scenarios such as credit card transactions,mobile banking,and cross-border payments.The study also highlights future directions,including the deployment of lightweight models and the integration of multimodal data for scalable fraud analytics.展开更多
This study explores ethical violations and audit failures in various large-scale corporate fraud cases.Using a qualitative descriptive method and based on secondary data from documented fraud cases and published audit...This study explores ethical violations and audit failures in various large-scale corporate fraud cases.Using a qualitative descriptive method and based on secondary data from documented fraud cases and published audit reports,the study applies the Fraud Triangle framework,focusing on how weak integrity,objectivity,and professional competence have undermined public trust in the auditing profession.Using a qualitative descriptive method and the Fraud Triangle framework,which includes pressure,opportunity,and rationalization,the study analyzes cases from Indonesia(SNP Finance,Jiwasraya),China(Evergrande),and Germany(Wirecard).The analysis reveals that many audit failures observed in this study appear to stem more from ethical challenges than from technical incapability,driven by client pressure,weak internal controls,and compromised auditor independence.These cases demonstrate a recurring global pattern in which auditors neglect their responsibility to act in the public interest.展开更多
This study explores ethical violations and audit failures in various large-scale corporate fraud cases.Using a qualitative descriptive method and based on secondary data from documented fraud cases and published audit...This study explores ethical violations and audit failures in various large-scale corporate fraud cases.Using a qualitative descriptive method and based on secondary data from documented fraud cases and published audit reports,the study applies the Fraud Triangle framework,focusing on how weak integrity,objectivity,and professional competence have undermined public trust in the auditing profession.Using a qualitative descriptive method and the Fraud Triangle framework,which includes pressure,opportunity,and rationalization,the study analyzes cases from Indonesia(SNP Finance,Jiwasraya),China(Evergrande),and Germany(Wirecard).The analysis reveals that many audit failures observed in this study appear to stem more from ethical challenges than from technical incapability,driven by client pressure,weak internal controls,and compromised auditor independence.These cases demonstrate a recurring global pattern in which auditors neglect their responsibility to act in the public interest.展开更多
Credit card fraud is one of the primary sources of operational risk in banks,and accurate prediction of fraudulent credit card transactions is essential to minimize banks’economic losses.Two key issues are faced in c...Credit card fraud is one of the primary sources of operational risk in banks,and accurate prediction of fraudulent credit card transactions is essential to minimize banks’economic losses.Two key issues are faced in credit card fraud detection research,i.e.,data category imbalance and data drift.However,the oversampling algorithm used in current research suffers from excessive noise,and the Long Short-Term Memory Network(LSTM)based temporal model suffers from gradient dispersion,which can lead to loss of model performance.To address the above problems,a credit card fraud detection method based on Random Forest-Wasserstein Generative Adversarial NetworkTemporal Convolutional Network(RF-WGAN-TCN)is proposed.First,the credit card data is preprocessed,the feature importance scores are calculated by Random Forest(RF),the features with lower importance are eliminated,and then the remaining features are standardized.Second,the Wasserstein Distance Improvement Generative Adversarial Network(GAN)is introduced to construct the Wasserstein Generative Adversarial Network(WGAN),the preprocessed data is input into the WGAN,and under the mutual game training of generator and discriminator,the fraud samples that meet the target distribution are obtained.Finally,the temporal convolutional network(TCN)is utilized to extract the long-time dependencies,and the classification results are output through the Softmax layer.Experimental results on the European cardholder dataset show that the method proposed in the paper achieves 91.96%,98.22%,and 81.95%in F1-Score,Area Under Curve(AUC),and Area Under the Precision-Recall Curve(AUPRC)metrics,respectively,and has higher prediction accuracy and classification performance compared with existing mainstream methods.展开更多
Purpose:This study aims to integrate large language models(LLMs)with interpretable machine learning methods to develop a multimodal data-driven framework for predicting corporate financial fraud,addressing the limitat...Purpose:This study aims to integrate large language models(LLMs)with interpretable machine learning methods to develop a multimodal data-driven framework for predicting corporate financial fraud,addressing the limitations of traditional approaches in long-text semantic parsing,model interpretability,and multisource data fusion,thereby providing regulatory agencies with intelligent auditing tools.Design/methodology/approach:Analyzing 5,304 Chinese listed firms’annual reports(2015-2020)from the CSMAD database,this study leverages the Doubao LLMs to generate chunked summaries and 256-dimensional semantic vectors,developing textual semantic features.It integrates 19 financial indicators,11 governance metrics,and linguistic characteristics(tone,readability)with fraud prediction models optimized through a group of Gradient Boosted Decision Tree(GBDT)algorithms.SHAP value analysis in the final model reveals the risk transmission mechanism by quantifying the marginal impacts of financial,governance,and textual features on fraud likelihood.Findings:The study found that LLMs effectively distill lengthy annual reports into semantic summaries,while GBDT algorithms(AUC>0.850)outperform the traditional Logistic Regression model in fraud detection.Multimodal fusion improved performance by 7.4%,with financial,governance,and textual features providing complementary signals.SHAP analysis revealed financial distress,governance conflicts,and narrative patterns(e.g.,tone anchoring,semantic thresholds)as key fraud indicators,highlighting managerial intent in report language.Research limitations:This study identifies three key limitations:1)lack of interpretability for semantic features,2)absence of granular fraud-type differentiation,and 3)unexplored comparative validation with other deep learning methods.Future research will address these gaps to enhance fraud detection precision and model transparency.Practical implications:The developed semantic-enhanced evaluation model provides a quantitative tool for assessing listed companies’information disclosure quality and enables practical implementation through its derivative real-time monitoring system.This advancement significantly strengthens capital market risk early warning capabilities,offering actionable insights for securities regulation.Originality/value:This study presents three key innovations:1)A novel“chunking-summarizationembedding”framework for efficient semantic compression of lengthy annual reports(30,000 words);2)Demonstration of LLMs’superior performance in financial text analysis,outperforming traditional methods by 19.3%;3)A novel“language-psychology-behavior”triad model for analyzing managerial fraud motives.展开更多
The task of detecting fraud in credit card transactions is crucial to ensure the security and stability of a financial system,as well as to enforce customer confidence in digital payment systems.Historically,credit ca...The task of detecting fraud in credit card transactions is crucial to ensure the security and stability of a financial system,as well as to enforce customer confidence in digital payment systems.Historically,credit card companies have used rulebased approaches to detect fraudulent transactions,but these have proven inadequate due to the complexity of fraud strategies and have been replaced by much more powerful solutions based on machine learning or deep learning algorithms.Despite significant progress,the current approaches to fraud detection suffer from a number of limitations:for example,it is unclear whether some transaction features are more effective than others in discriminating fraudulent transactions,and they often neglect possible correlations among transactions,even though they could reveal illicit behaviour.In this paper,we propose a novel credit card fraud detection(CCFD)method based on a transaction behaviour-based hierarchical gated network.First,we introduce a feature-oriented extraction module capable of identifying key features from original transactions,and such analysis is effective in revealing the behavioural characteristics of fraudsters.Second,we design a transaction-oriented extraction module capable of capturing the correlation between users’historical and current transactional behaviour.Such information is crucial for revealing users’sequential behaviour patterns.Our approach,called transactional-behaviour-based hierarchical gated network model(TbHGN),extracts two types of new transactional features,which are then combined in a feature interaction module to learn the final transactional representations used for CCFD.We have conducted extensive experiments on a real-world credit card transaction dataset with an increase in average F1 between 1.42%and 6.53%and an improvement in average AUC between 0.63%and 2.78%over the state of the art.展开更多
银行账户欺诈检测是安全行业中的一大难题,主要因为欺诈模式的快速变化和合法账户与欺诈账户之间的显著数据不平衡问题.传统检测方法在一定程度上能解决这一问题,但常常面临较高的误报率,并且在应对新型欺诈行为时表现较差.本文提出了...银行账户欺诈检测是安全行业中的一大难题,主要因为欺诈模式的快速变化和合法账户与欺诈账户之间的显著数据不平衡问题.传统检测方法在一定程度上能解决这一问题,但常常面临较高的误报率,并且在应对新型欺诈行为时表现较差.本文提出了一种创新方法,将专家模型(Mixture of Experts,MoE)与基于深度神经网络的少数类过采样技术(DNN-SMOTE)相结合,以提高银行账户欺诈检测的效果.MoE模型通过多个专门训练的子模型捕获不同类型的欺诈行为特征,而DNN-SMOTE则通过生成高质量的少数类合成样本,显著缓解了类别不平衡的问题.在一个公开的银行账户欺诈数据集上,实验结果表明该方法的分类准确率达到了97.38%,真阳性准确率为87.02%.这表明所提出的模型在检测欺诈账户和合法账户之间具有良好的平衡性能.这些结果验证了MoE与DNN-SMOTE结合的有效性,为实际场景中的银行账户欺诈检测提供了一个强健且高效的解决方案.展开更多
基金supported by the National Research Foundation of Korea(NRF)funded by the Korea government(RS-2023-00249743)Additionally,this research was supported by the Global-Learning&Academic Research Institution for Master’s,PhD Students,and Postdocs(LAMP)Program of the National Research Foundation of Korea(NRF)grant funded by the Ministry of Education(RS-2024-00443714)This research was also supported by the“Research Base Construction Fund Support Program”funded by Jeonbuk National University in 2025.
文摘Traditional anomaly detection methods often assume that data points are independent or exhibit regularly structured relationships,as in Euclidean data such as time series or image grids.However,real-world data frequently involve irregular,interconnected structures,requiring a shift toward non-Euclidean approaches.This study introduces a novel anomaly detection framework designed to handle non-Euclidean data by modeling transactions as graph signals.By leveraging graph convolution filters,we extract meaningful connection strengths that capture relational dependencies often overlooked in traditional methods.Utilizing the Graph Convolutional Networks(GCN)framework,we integrate graph-based embeddings with conventional anomaly detection models,enhancing performance through relational insights.Ourmethod is validated on European credit card transaction data,demonstrating its effectiveness in detecting fraudulent transactions,particularly thosewith subtle patterns that evade traditional,amountbased detection techniques.The results highlight the advantages of incorporating temporal and structural dependencies into fraud detection,showcasing the robustness and applicability of our approach in complex,real-world scenarios.
文摘This study investigates the dark side of the non-fungible token(NFT)marketplace,with a focus on understanding the risks,and underlying factors driving fraud in the NFT ecosystem.Using the fraud triangle framework,this study examines pressure,opportunity,and rationalization from individual and organizational perspectives.The research provides a comprehensive understanding of the contributing factors to NFT marketplace fraud by analyzing the reasons behind fraudulent actions.A conceptual framework is developed that includes ten propositions to aid in understanding the complexity of this issue.This study’s outcomes will assist policymakers in crafting efficient approaches to mitigate fraud within the NFT marketplace.
文摘This research aims to understand the causes of fraud through the approach of the Fraud Triangle Theory,which includes three main factors:pressure,opportunity,rationalization.By using the Systematic Literature Review method from various relevant international journals,it analyzed systematically to identify patterns,trends,and theoretical contributions to efforts in detecting and preventing fraud.The results of this study show that the three factors in the Fraud Triangle Theory significantly contribute to the occurrence of fraud.Opportunity as the most dominant factor is caused by weak internal control systems and lack of oversight.In addition,economic pressure,and a permissive organizational environment,as well as the rationalization processes by individuals,also increase the tendency for a person to commit fraud.These findings emphasize the need for a comprehensive approach in fraud prevention strategies,through ethical values,the establishment of an organizational culture with integrity,and the implementation of more effective internal control and oversight.
基金supported by project Finance for all(F4A),funded by the"Institución Gran Duque de Alba"and"Diputación provincial deávila"under the grant 3364/2022.
文摘Reducing the risk of fraud in credit card transactions is crucial for the competitiveness of companies,especially in Latin American countries.This study aims to establish measures for preventing and detecting fraud in the use of credit cards in shops through analytical methods(data mining,machine learning and artificial intelligence).To achieve this objective,the study employs a predictive methodology using descriptive and exploratory statistics and frequency,frequency&monetary(RFM)classification techniques,differentiating between SMEs and large businesses via cluster analysis and supervised models.A dataset of 221,292 card records from a Latin American merchant payment gateway for the year 2022 is used.For fraud alerts,the classification model has been selected for small and medium–sized merchants,and the multilayer perceptron(MLP)neural network has been selected for large merchants.Random forest or Gini decision tree models have been selected as backup models for retraining.For the detection of punctual fraud patterns,the K-means and partitioning around medoids(PAM)models have been selected,depending on the type of trade.The results revealed that the application of the identified models would have prevented between 48 and 85%of fraud transactions,depending on the trade size.Despite the promising results,continuous updating is recommended,as fraudsters frequently implement new fraud techniques.
基金funded by the Deanship of Scientific Research,Vice Presidency for Graduate Studies and Scientific Research,King Faisal University,Saudi Arabia[Grant No.KFU241683].
文摘This paper proposes a novel hybrid fraud detection framework that integrates multi-stage feature selection,unsupervised clustering,and ensemble learning to improve classification performance in financial transaction monitoring systems.The framework is structured into three core layers:(1)feature selection using Recursive Feature Elimination(RFE),Principal Component Analysis(PCA),and Mutual Information(MI)to reduce dimensionality and enhance input relevance;(2)anomaly detection through unsupervised clustering using K-Means,Density-Based Spatial Clustering(DBSCAN),and Hierarchical Clustering to flag suspicious patterns in unlabeled data;and(3)final classification using a voting-based hybrid ensemble of Support Vector Machine(SVM),Random Forest(RF),and Gradient Boosting Classifier(GBC).The experimental evaluation is conducted on a synthetically generated dataset comprising one million financial transactions,with 5% labelled as fraudulent,simulating realistic fraud rates and behavioural features,including transaction time,origin,amount,and geo-location.The proposed model demonstrated a significant improvement over baseline classifiers,achieving an accuracy of 99%,a precision of 99%,a recall of 97%,and an F1-score of 99%.Compared to individual models,it yielded a 9% gain in overall detection accuracy.It reduced the false positive rate to below 3.5%,thereby minimising the operational costs associated with manually reviewing false alerts.The model’s interpretability is enhanced by the integration of Shapley Additive Explanations(SHAP)values for feature importance,supporting transparency and regulatory auditability.These results affirm the practical relevance of the proposed system for deployment in real-time fraud detection scenarios such as credit card transactions,mobile banking,and cross-border payments.The study also highlights future directions,including the deployment of lightweight models and the integration of multimodal data for scalable fraud analytics.
文摘This study explores ethical violations and audit failures in various large-scale corporate fraud cases.Using a qualitative descriptive method and based on secondary data from documented fraud cases and published audit reports,the study applies the Fraud Triangle framework,focusing on how weak integrity,objectivity,and professional competence have undermined public trust in the auditing profession.Using a qualitative descriptive method and the Fraud Triangle framework,which includes pressure,opportunity,and rationalization,the study analyzes cases from Indonesia(SNP Finance,Jiwasraya),China(Evergrande),and Germany(Wirecard).The analysis reveals that many audit failures observed in this study appear to stem more from ethical challenges than from technical incapability,driven by client pressure,weak internal controls,and compromised auditor independence.These cases demonstrate a recurring global pattern in which auditors neglect their responsibility to act in the public interest.
文摘This study explores ethical violations and audit failures in various large-scale corporate fraud cases.Using a qualitative descriptive method and based on secondary data from documented fraud cases and published audit reports,the study applies the Fraud Triangle framework,focusing on how weak integrity,objectivity,and professional competence have undermined public trust in the auditing profession.Using a qualitative descriptive method and the Fraud Triangle framework,which includes pressure,opportunity,and rationalization,the study analyzes cases from Indonesia(SNP Finance,Jiwasraya),China(Evergrande),and Germany(Wirecard).The analysis reveals that many audit failures observed in this study appear to stem more from ethical challenges than from technical incapability,driven by client pressure,weak internal controls,and compromised auditor independence.These cases demonstrate a recurring global pattern in which auditors neglect their responsibility to act in the public interest.
基金supported by the National Natural Science Foundation of China under Grant No.62466001the Talent Plan Project of Fuzhou City of Jiangxi Province of China under the Grant No.2021ED008+1 种基金the Opening Project of Jiangxi Key Laboratory of Cybersecurity Intelligent Perception under the Grant No.JKLCIP202202the Priority Unveiled Marshalling Project of Fuzhou City of Jiangxi Province of China under the Grant No.2023JBB026.
文摘Credit card fraud is one of the primary sources of operational risk in banks,and accurate prediction of fraudulent credit card transactions is essential to minimize banks’economic losses.Two key issues are faced in credit card fraud detection research,i.e.,data category imbalance and data drift.However,the oversampling algorithm used in current research suffers from excessive noise,and the Long Short-Term Memory Network(LSTM)based temporal model suffers from gradient dispersion,which can lead to loss of model performance.To address the above problems,a credit card fraud detection method based on Random Forest-Wasserstein Generative Adversarial NetworkTemporal Convolutional Network(RF-WGAN-TCN)is proposed.First,the credit card data is preprocessed,the feature importance scores are calculated by Random Forest(RF),the features with lower importance are eliminated,and then the remaining features are standardized.Second,the Wasserstein Distance Improvement Generative Adversarial Network(GAN)is introduced to construct the Wasserstein Generative Adversarial Network(WGAN),the preprocessed data is input into the WGAN,and under the mutual game training of generator and discriminator,the fraud samples that meet the target distribution are obtained.Finally,the temporal convolutional network(TCN)is utilized to extract the long-time dependencies,and the classification results are output through the Softmax layer.Experimental results on the European cardholder dataset show that the method proposed in the paper achieves 91.96%,98.22%,and 81.95%in F1-Score,Area Under Curve(AUC),and Area Under the Precision-Recall Curve(AUPRC)metrics,respectively,and has higher prediction accuracy and classification performance compared with existing mainstream methods.
基金supported by the 2021 Guangdong Province(China)Science and Technology Plan Project“Research and Application of Key Technologies for Multi-level Knowledge Retrieval Based on Big Data Intelligence”(Project No.2021B0101420004)the 2022 commissioned project“Cross-border E-commerce Taxation and Related Research”from the State Taxation Administration Guangdong Provincial Taxation Bureau,China.
文摘Purpose:This study aims to integrate large language models(LLMs)with interpretable machine learning methods to develop a multimodal data-driven framework for predicting corporate financial fraud,addressing the limitations of traditional approaches in long-text semantic parsing,model interpretability,and multisource data fusion,thereby providing regulatory agencies with intelligent auditing tools.Design/methodology/approach:Analyzing 5,304 Chinese listed firms’annual reports(2015-2020)from the CSMAD database,this study leverages the Doubao LLMs to generate chunked summaries and 256-dimensional semantic vectors,developing textual semantic features.It integrates 19 financial indicators,11 governance metrics,and linguistic characteristics(tone,readability)with fraud prediction models optimized through a group of Gradient Boosted Decision Tree(GBDT)algorithms.SHAP value analysis in the final model reveals the risk transmission mechanism by quantifying the marginal impacts of financial,governance,and textual features on fraud likelihood.Findings:The study found that LLMs effectively distill lengthy annual reports into semantic summaries,while GBDT algorithms(AUC>0.850)outperform the traditional Logistic Regression model in fraud detection.Multimodal fusion improved performance by 7.4%,with financial,governance,and textual features providing complementary signals.SHAP analysis revealed financial distress,governance conflicts,and narrative patterns(e.g.,tone anchoring,semantic thresholds)as key fraud indicators,highlighting managerial intent in report language.Research limitations:This study identifies three key limitations:1)lack of interpretability for semantic features,2)absence of granular fraud-type differentiation,and 3)unexplored comparative validation with other deep learning methods.Future research will address these gaps to enhance fraud detection precision and model transparency.Practical implications:The developed semantic-enhanced evaluation model provides a quantitative tool for assessing listed companies’information disclosure quality and enables practical implementation through its derivative real-time monitoring system.This advancement significantly strengthens capital market risk early warning capabilities,offering actionable insights for securities regulation.Originality/value:This study presents three key innovations:1)A novel“chunking-summarizationembedding”framework for efficient semantic compression of lengthy annual reports(30,000 words);2)Demonstration of LLMs’superior performance in financial text analysis,outperforming traditional methods by 19.3%;3)A novel“language-psychology-behavior”triad model for analyzing managerial fraud motives.
基金supported in part by the National Natural Science Foundation of China(61972241)the Natural Science Foundation of Shanghai(24ZR1427500,22ZR1427100)+1 种基金the Key Projects of Natural Science Research in Anhui Higher Education Institutions(2022AH051909)Bengbu University 2021 High-Level Scientific Research and Cultivation Project(2021pyxm04).
文摘The task of detecting fraud in credit card transactions is crucial to ensure the security and stability of a financial system,as well as to enforce customer confidence in digital payment systems.Historically,credit card companies have used rulebased approaches to detect fraudulent transactions,but these have proven inadequate due to the complexity of fraud strategies and have been replaced by much more powerful solutions based on machine learning or deep learning algorithms.Despite significant progress,the current approaches to fraud detection suffer from a number of limitations:for example,it is unclear whether some transaction features are more effective than others in discriminating fraudulent transactions,and they often neglect possible correlations among transactions,even though they could reveal illicit behaviour.In this paper,we propose a novel credit card fraud detection(CCFD)method based on a transaction behaviour-based hierarchical gated network.First,we introduce a feature-oriented extraction module capable of identifying key features from original transactions,and such analysis is effective in revealing the behavioural characteristics of fraudsters.Second,we design a transaction-oriented extraction module capable of capturing the correlation between users’historical and current transactional behaviour.Such information is crucial for revealing users’sequential behaviour patterns.Our approach,called transactional-behaviour-based hierarchical gated network model(TbHGN),extracts two types of new transactional features,which are then combined in a feature interaction module to learn the final transactional representations used for CCFD.We have conducted extensive experiments on a real-world credit card transaction dataset with an increase in average F1 between 1.42%and 6.53%and an improvement in average AUC between 0.63%and 2.78%over the state of the art.
文摘银行账户欺诈检测是安全行业中的一大难题,主要因为欺诈模式的快速变化和合法账户与欺诈账户之间的显著数据不平衡问题.传统检测方法在一定程度上能解决这一问题,但常常面临较高的误报率,并且在应对新型欺诈行为时表现较差.本文提出了一种创新方法,将专家模型(Mixture of Experts,MoE)与基于深度神经网络的少数类过采样技术(DNN-SMOTE)相结合,以提高银行账户欺诈检测的效果.MoE模型通过多个专门训练的子模型捕获不同类型的欺诈行为特征,而DNN-SMOTE则通过生成高质量的少数类合成样本,显著缓解了类别不平衡的问题.在一个公开的银行账户欺诈数据集上,实验结果表明该方法的分类准确率达到了97.38%,真阳性准确率为87.02%.这表明所提出的模型在检测欺诈账户和合法账户之间具有良好的平衡性能.这些结果验证了MoE与DNN-SMOTE结合的有效性,为实际场景中的银行账户欺诈检测提供了一个强健且高效的解决方案.
