TY - JOUR

T1 - Fraud detection in capital markets

T2 - A novel machine learning approach

AU - Yi, Ziwei

AU - Cao, Xinwei

AU - Pu, Xujin

AU - Wu, Yiding

AU - Chen, Zuyan

AU - Khan, Ameer Tamoor

AU - Francis, Adam

AU - Li, Shuai

N1 - Publisher Copyright: © 2023 Elsevier Ltd

PY - 2023

Y1 - 2023

N2 - Traditional auditing methods require collating massive amounts of financial indicators and related transaction data, which can be labor-intensive. Typical machine learning models are relatively weak for imbalanced data, and this work aims to focus on a novel approach to fraud detection. This paper presents a fraud detection framework via adopting a machine learning method integrated with a recently proposed meta-heuristics algorithm Egret Swarm Optimization Algorithm (ESOA). A cost-sensitive objective function and loss function were then constructed, and a non-linear model was used to map the predicted values into the labels of 0 (non-fraud) and 1 (fraud). In the experiment section, an AAER benchmark dataset collected by the UCB's Center for Financial Reporting and Management is utilized to verify the performance of the proposed approach. A detailed comparison with recently proposed state-of-the-art algorithms such as Logit (67.20%), SVM-FK (62.60%), RUSBoost (72.60%), as well as BAS (84.90%) indicates that ESOA (96.27%) outperforms the other algorithms in terms of Accuracy (ACC), Sensitivity (SEN), Precision (PREC), and Area Under the Curve (AUC) metrics. To our knowledge, this is the highest fraud detection accuracy reported in the existing literature.

AB - Traditional auditing methods require collating massive amounts of financial indicators and related transaction data, which can be labor-intensive. Typical machine learning models are relatively weak for imbalanced data, and this work aims to focus on a novel approach to fraud detection. This paper presents a fraud detection framework via adopting a machine learning method integrated with a recently proposed meta-heuristics algorithm Egret Swarm Optimization Algorithm (ESOA). A cost-sensitive objective function and loss function were then constructed, and a non-linear model was used to map the predicted values into the labels of 0 (non-fraud) and 1 (fraud). In the experiment section, an AAER benchmark dataset collected by the UCB's Center for Financial Reporting and Management is utilized to verify the performance of the proposed approach. A detailed comparison with recently proposed state-of-the-art algorithms such as Logit (67.20%), SVM-FK (62.60%), RUSBoost (72.60%), as well as BAS (84.90%) indicates that ESOA (96.27%) outperforms the other algorithms in terms of Accuracy (ACC), Sensitivity (SEN), Precision (PREC), and Area Under the Curve (AUC) metrics. To our knowledge, this is the highest fraud detection accuracy reported in the existing literature.

KW - Egret Swarm Optimization Algorithm

KW - ESOA

KW - Fraud detection

KW - Listed corporates

KW - Machine learning

U2 - 10.1016/j.eswa.2023.120760

DO - 10.1016/j.eswa.2023.120760

M3 - Journal article

AN - SCOPUS:85162084094

VL - 231

JO - Expert Systems with Applications

JF - Expert Systems with Applications

SN - 0957-4174

M1 - 120760

ER -