Jul 2026· 2026 International Conference on Intelligent and Sustainable AI Systems (ICOSAAS)· pp. 1124-1130· 0 citations· 22 references
Abstract
The rise of digital payments has magnified credit card fraud issues the complexity, scope and therefore the attack strategies have evolved to become an enormous obstacle for the traditional, static (rule based) and single- modal methods of ML in learning dynamic transactional patterns. In this paper we introduced a multimodal fraud detection approach combining the typical numerical features characterizing single transactional entities with their time- series transaction behaviour so that we could achieve higher accuracy in identifying frauds. The former represents properties for each transaction; temporal patterns in user spending are captured by a Gated Recurrent Unit (GRU), a network effectively modelling sequences of events (purchase history in our context) and then combined to finally perform binary classification in which either fraudulent or real transaction class is detected. We use the public domain 284,807 transaction records from credit card dataset; fraud transactions account for 0.172 of the samples. The experiment shows our novel framework (multimodal GRU) outperformed existing ML algorithms and single-mode techniques, having the recall reach 0.89, F1-score value 0.90 and AUC score 0.99. It’s highly scalable and efficient to use in real time monitoring applications. Our proposed method of combining two types of features: a set of single number characteristics and sequence-based information to successfully solve this challenging problem in modern finance.
The rapid evolution of digital banking and e-commerce puts online payment fraud problems at another level. Because of the large class imbalance present, coupled with the ever-changing nature of fraud, in the instance of credit card fraud, it becomes doubly challenging to achieve real-time detection. Fraudsters continuously evolve and modify their strategies to capture the weaknesses in the digital banking system. To contribute to this problem, this paper provides an approach to detect credit card fraud in the context of machine learning, with a comprehensive set of data preparation methods, the Synthetic Minority Over-sampling Technique (SMOTE) for class imbalance, normalization, feature selection by the information gain, and a hybrid approach using two different supervised learning techniques. The study is based on a readily accessible dataset of European credit card transactions from September 2013, which is composed of 284,807 transactions, of which only 492 (0.172%) are considered to belong to the fraudulent class. In this approach, a Support Vector Machine (SVM) is used to define the critical decision boundaries, and probabilistic outputs from the SVM are used as input to train an artificial neural network (ANN). This hybrid SVM-ANN approach leverages the strengths of both models and improves the classification performance using margin-based learning and nonlinear representation technologies. The study tackles some of the major issues of fraud detection, such as the use of feature sets transformed by PCA for confidentiality, the evaluation of metrics for models that capture rare events, and the generalization of models to be used in imbalanced situations. The proposed approach performs most other traditional classifiers, as evidenced by the experiments and the stated evaluation metrics, which include accuracy, precision, F-measure, and recall. The results show how important it is to use hybrid modeling and strong preprocessing techniques to build scalable and reliable systems for credit card fraud detection.
Firdos R. Kachur, Shivangi Agarwal, Gargi Phadke· International Conference Com...· 0 citations
For identifying fraud within a vast array of digitally conducted financial transactions, it is imperative for financial transaction systems to be scalable and accurate. One of the common uses of machine learning in the financial sector is credit card fraud detection, particularly when manual monitoring of the massive volume of financial transactions is impractical. The proposed system is designed using the XGBoost classifier. This approach improves the ability to detect fraudulent credit card transactions through various behavioral and geographical aspects, including transaction time, transaction frequency, and the distance between the cardholder and merchant locations. Experimental results indicate that the proposed system is capable of detecting fraudulent transactions with a precision of approximately 99.8% and a high ROC-AUC score. The scalable framework designed in this study can assist financial organizations in improving the security and effectiveness of fraud detection systems.
J. Kumar, J. Vasu, Kalagadda Kiran Kumar· 2026 6th International Confe...· 0 citations
The increasing use of online financial services, including online banking, mobile payment platforms, UPI transactions, and electronic fund transfers, has significantly increased the risk of financial fraud in today's digital economy. Conventional rule-based fraud detection approaches are often ineffective in handling continuously changing fraud techniques, leading to considerable losses for both customers and financial organizations. To resolve these limitations, this work presents a Hybrid Financial_Fraud Detection System that combines multiple machine_learning techniques to achieve higher_detection accuracy and enable real-time transaction analysis. The suggested framework employs a stacked ensemble approach in which Logistic_Regression, Decision_Tree, and XGBoost act as base models, while a meta-classifier produces the final fraud prediction. The model evaluates several transaction-related parameters, including transaction amount, account balance, transaction frequency, average transaction value, account age, previous fraud history, transaction time, and indicators of foreign transactions. Customer details are automatically obtained from a customer database, allowing the system to perform transaction-specific and personalized risk assessment. A banking-oriented application_interface has been developed to enable users to carry out transactions through different payment channels such as UPI, bank_transfer, debit_card, credit_card, NEFT, RTGS, and IMPS. Each transaction is processed and analyzed instantly, assigned a fraud risk score, and categorized as either genuine or fraudulent. In addition, the system records transaction history, updates account balances, and provides a monitoring dashboard to support continuous fraud analysis. By integrating ensemble learning_techniques with behavioral transaction analysis and real-time monitoring, the proposed system improves the overall effectiveness of fraud detection. The framework is designed to strengthen transaction security, minimize financial_losses, and offer a scalable and intelligent solution for fraud prevention in modern digital financial environments.
G S Rahul Gorpade, Dr. Ashwin M· International Research Journ...· 0 citations
A comprehensive, real-time credit card fraud detection framework that integrates geospatial and occupational user behavior features with supervised machine learning models is proposed and successfully bridges the gap between theoretical machine learning models and practical, low-latency deployment in financial technologies.
Narayana Reddy T, Arun Kumar T M, Nagaraju V et al.· American Journal of AI Cyber...· 0 citations
The rapid growth of digital banking, online transactions, and the electronic payment of the systems has been
significantly increased the use of credit cards, results in the parallel rise in fraudulent activities. Traditional methods of
detecting the fraud are often unable to handle large-scale transaction data and evolving fraud patterns efficiently. This paper
presents an AI-Based Credit Card Fraud Detection System using Machine Learning to identify suspicious transactions
accurately and in the real time. To identify patterns and detect potential fraud, the proposed system analyzes past
transaction data containing various transaction-related attributes amount, merchant details, category, time, and customer
information. Data preprocessing techniques including K-Nearest Neighbor (KNN) imputation are applied to handle missing
values, followed by label encoding and Z-score normalization to improve data quality and consistency. Feature engineering
is performed to extract meaningful transaction patterns. The processed data is then classified using the XGBoost algorithm,
which improves prediction accuracy through sequential learning and optimized decision trees. Experimental evaluation is
carried out using a credit card transaction dataset containing legitimate and fraudulent records.
P. Ravikumar, Gowrav A. S., A. C. N. et al.· International Journal of Inn...· 0 citations
Financial fraud has emerged as one of the most significant challenges in the modern digital economy due to the rapid growth of online banking, mobile payments, electronic commerce, and digital financial services. Traditional fraud detection systems primarily rely on predefined rules and expert-generated patterns; however, such approaches often fail to identify newly emerging fraud techniques and sophisticated fraudulent behaviors. Consequently, there is an increasing need for intelligent and adaptive fraud detection systems capable of learning from historical transaction data and identifying suspicious activities with high accuracy.
This research presents an intelligent machine learning-based approach for fraud detection in financial transactions. The study investigates and compares the effectiveness of multiple machine learning algorithms, including Logistic Regression, Decision Tree, Random Forest, Extreme Gradient Boosting (XGBoost), and Artificial Neural Networks (ANN), in detecting fraudulent financial activities. The proposed methodology incorporates data preprocessing, feature engineering, handling class imbalance, model training, and comprehensive performance evaluation.
A publicly available credit card transaction dataset is utilized for experimental analysis. Data preprocessing techniques such as normalization, missing value handling, and feature selection are applied to improve model performance. Since fraud datasets are typically highly imbalanced, class balancing techniques are incorporated to enhance the detection of minority fraudulent transactions.
The performance of the selected algorithms is evaluated using Accuracy, Precision, Recall, F1-Score, and Area Under the Receiver Operating Characteristic Curve (AUC-ROC).
Comparative analysis demonstrates that ensemble learning methods, particularly Random Forest and XGBoost, outperform traditional classification techniques by achieving higher detection rates and lower false positive rates. Furthermore, the study highlights the practical applicability of machine learning models in real-world financial environments where rapid and accurate fraud detection is essential.
The findings of this research contribute to the growing field of financial fraud analytics by providing a comparative framework for evaluating machine learning algorithms under consistent experimental conditions. The proposed approach offers valuable insights for financial institutions, banking organizations, and cybersecurity professionals seeking to enhance fraud prevention systems and minimize financial losses.
Keywords: Fraud Detection, Machine Learning, Financial Transactions, Random Forest, XGBoost, Artificial Neural Networks, Classification, Financial Security.
Dr. Abdul Majid Farooqi Dr. Abdul Majid Farooqi, Laiba Khan LAIBA KHAN· International Scientific Jou...· 0 citations