Jun 2026· International Journal for Research in Applied Science and Engineering Technology· Vol 14, pp. 2806-2810· 0 citations
TL;DR
FraudX is presented, an end-to-end explainable fraud detection framework that combines supervised learning with rule-based categorization and human-readable explanations and demonstrates how operational thresholding, featureimportance-driven explanations, and analyst-oriented visualization can be integrated into a practical fraud triage workflow suitable for academic demonstration and prototype deployment.
Abstract
Financial fraud causes significant economic loss and erodes trust in digital payment ecosystems. Traditional rule-only
systems struggle with evolving attack patterns, while pure black-box machine learning models are difficult for analysts to
interpret during investigations. This paper presents **FraudX**, an end-to-end explainable fraud detection framework that
combines supervised learning with rule-based categorization and human-readable explanations. The system generates a labeled
synthetic transaction dataset of 10,000 records with 20+ behavioral and contextual features, trains a Random Forest classifier
with standardized preprocessing, and deploys inference through an interactive Streamlit dashboard. For each transaction,
FraudX outputs a probability-based *Suspicion Score*, a threshold-controlled suspicious flag, a prioritized *Fraud_Type* label,
and concise *Suspicion_Reasons*. Experimental evaluation on held-out synthetic test data reports 95.2% accuracy, 94.8%
precision, 93.5% recall, and 94.1% F1-score. The proposed architecture demonstrates how operational thresholding, featureimportance-driven explanations, and analyst-oriented visualization can be integrated into a practical fraud triage workflow
suitable for academic demonstration and prototype deployment.
The rapid growth of digital financial services has significantly increased the volume of online transactions, making fraud detection a critical challenge for financial institutions. Traditional machine learning models often provide strong predictive performance but lack interpretability, limiting trust and practical adoption in financial decision-making. This study proposes an Explainable Artificial Intelligence (XAI)-based fraud detection framework for FinTech transactions using the Kaggle Credit Card Fraud Detection dataset containing 284,807 transactions, including 492 fraudulent cases. To address severe class imbalance, Synthetic Minority Oversampling Technique (SMOTE) was applied, increasing the dataset to 568,630 balanced instances. Data preprocessing involved feature scaling and train–test splitting prior to model training. Three machine learning algorithms—Logistic Regression, Random Forest, and Extreme Gradient Boosting (XGBoost) were developed and evaluated using accuracy, precision, recall, F1-score, and ROC-AUC metrics.
The experimental results demonstrate strong predictive performance across all models. Logistic Regression achieved 94.50% accuracy, 97.32% precision, 91.51% recall, 94.33% F1-score, and a ROC-AUC of 94.50%. Random Forest produced the highest overall performance with 99.99% accuracy, 99.98% precision, 100.00% recall, 99.99% F1-score, and 99.99% ROC-AUC. XGBoost also achieved excellent results with 99.97% accuracy, 99.94% precision, 100.00% recall, 99.97% F1-score, and 99.97% ROC-AUC. To improve model transparency, SHapley Additive exPlanations (SHAP) and Local Interpretable Model-Agnostic Explanations (LIME) were integrated with the XGBoost model to provide both global and local interpretability. SHAP analysis identified transaction amount and several transformed principal component features as the most influential predictors of fraudulent behavior, while LIME provided instance-level explanations for individual fraud predictions. Feature importance analysis from Random Forest and XGBoost further validated the consistency of the most influential variables.
The findings demonstrate that combining high-performing machine learning models with explainable AI techniques can significantly enhance fraud detection accuracy while maintaining transparency and interpretability. The proposed framework offers a reliable and practical approach for intelligent fraud prevention in financial technology systems and supports trustworthy decision-making in real-world financial environments.
D. Aina, J. A. Ayeni, A. Ogunjobi et al.· International journal of res...· 0 citations
Financial fraud has become a major problem in modern digital transactions. Traditional fraud detection systems mainly focus on prediction accuracy, but they do not clearly explain why a transaction is marked as fraudulent. This creates trust and transparency issues in banking and financial applications. To solve this problem, this research paper proposes FraudXAI-Net, an explainability-driven intelligent fraud detection framework that combines machine learning and Explainable Artificial Intelligence (XAI). The proposed framework performs data preprocessing, feature scaling, class imbalance handling, model training, fraud prediction, and explainability analysis. Multiple machine learning models such as Logistic Regression, Random Forest, and XGBoost are used for fraud classification. SHAP (Shapley Additive Explanations) is applied to identify the contribution of each feature in prediction results. The framework improves both fraud detection performance and model interpretability. Experimental results show that the XGBoost model achieved the best performance with high accuracy, precision, recall, and F1-score. The SHAP-based explainability method successfully highlighted the important features responsible for fraudulent transactions. The proposed FraudXAI-Net framework can help financial institutions build more reliable, transparent, and intelligent fraud detection systems.
Jakkula Jayanthi· International Journal of Res...· 0 citations
Credit card fraud detection is widely recognized as a critical issue in digital financial systems due to the rapid growth of online transactions. However, existing approaches often fail to provide a comprehensive comparative framework and show limited effectiveness under highly imbalanced conditions. In this context, this study proposes a systematic machine learning pipeline to analyze fraudulent transactions using the IEEE-CIS dataset (590,000 records, 3–4% fraud). The approach integrates data preprocessing, feature encoding, and stratified sampling, followed by the evaluation of multiple supervised models including Logistic Regression, Decision Tree, Random Forest, SVM, Naive Bayes, and XGBoost. Performance is assessed using Accuracy, Precision, Recall, F1-score, and ROC-AUC. The results indicate that ensemble-based methods outperform traditional classifiers. In particular, XGBoost achieves the highest performance (accuracy 98.22%, ROC-AUC 0.9545, F1-score 0.6759), suggesting a strong ability to handle class imbalance. Nevertheless, no model fully optimizes both precision and recall, highlighting a key limitation in current fraud detection systems. Taken together, these findings suggest that ensemble learning provides a robust baseline, while further improvements through hybrid and cost-sensitive approaches are required for effective real-world deployment.
Financial fraud is moving very swiftly in today’s technologically advanced where everything is connected. This makes it extremely challenging for banks and other institutions of finance to follow the rules, preserve their customer trust, and run their businesses with integrity. Traditionally based on rules, detection systems can miss both small and big dangers when the number of transactions goes up and schemes for fraud get more intricate. Combining Machine Learning (ML) with Customer Relationship Management (CRM) data models is a powerful and versatile technique to stop fraud in this instance. Machine learning algorithms can identify hidden problems and anticipate fraud faster and more precisely by integrating information from CRM systems about past interactions with customers, behavior, and transactions in general. This work investigates a methodology that incorporates supervised and unsupervised methods of learning with enhanced CRM datasets in order to create sophisticated detection of fraud models. The methodology demonstrates data preprocessing, standardized feature engineering, and model training based on real financial parameters, including transaction frequency, alterations in typical customer behavior, and assessment of risk ratings. It also says that integrating CRM makes it less difficult for businesses to see the big picture of their customers, which enables these individuals to go from checking transactions by themselves to making decisions according to the situation. The suggested technique is to continually acquire knowledge and enhance the model so that it can keep up with the latest fraud strategies while minimizing the number of false positives that might adversely affect actual customers. This paper demonstrates the fact that banks may find fraud, minimize risks before they happen, and make their clients happy by combining ML and CRM standard data models together in an effective manner.
Satyendra Kumar Vanapalli· International Journal of Mac...· 0 citations
In this paper, a forensic end-to-end machine learning pipeline to detect financial fraud in PaySim synthetic mobile money dataset a corpus of 6,362,620 transactions with an extreme class imbalance ratio of 0.129% is presented. Proposed system is based on the distributed MLlib framework of Apache Spark in order to overcome the scalability issues associated with processing millions of financial records. Multicollinear raw ledger features are substituted with seven theoretically-based feature engineering transformations such as accounting discrepancy indices and account drain ratios as well as high-risk categorical flags. Ten classifiers such as Logistic Regression, Decision Tree variants, Random Forest ensembles, Gradient Boosted Trees, Linear SVC and a Factorization Machine are considered in a forensic model tournament. The imbalance between classes is alleviated through random undersampling to a balanced 50:50 training sample. The Matthews Correlation Coefficient (MCC) is the statistically rigorous evaluation index of the imbalanced binary classification that allows model selection. The winning ensemble model has a Recall of 99.39%, a Precision of 100.00%, an MCC of 0.9941 and a Financial Protection Index of over 96% which shows the operational reliability and good business value in preserving capital. Limitations regarding the synthetic nature of the data, the absence of sequence modeling, the schema dependence of the engineered features and the batch nature of the deployment are also discussed.
Anuran Baruah, Sahithya Sambhrama, K. S· 2026 4th International Confe...· 0 citations
Credit card fraud detection is challenging due to extreme class imbalance in transaction data and the evolving nature of fraudulent behavior, which causes traditional supervised classifiers to favor legitimate transactions and report misleadingly high accuracy. This paper presents a comparative study of four unsupervised anomaly detection methods trained exclusively on normal transactions: a reconstruction-based autoencoder, Isolation Forest, One-Class SVM, and Deep SVDD, evaluated on two highly imbalanced datasets, the Credit Card Fraud dataset and the PaySim simulated mobile money dataset. Continuous features are standardized, and each model generates anomaly scores that are converted to binary predictions using percentile-based thresholds at 90%, 95.5%, and 99%. Performance is assessed on stratified test sets using ROC-AUC, PR-AUC, precision, recall, F1-score, and accuracy to examine trade-offs between false positives and missed fraud cases. Results show that Deep SVDD achieves the highest precision, recall, and F1-score across both datasets, while the autoencoder also demonstrates strong and stable performance. Isolation Forest offers competitive results with lower computational cost but slightly reduced recall, whereas One-Class SVM exhibits greater sensitivity to threshold selection. These findings highlight the impact of threshold calibration, metric selection, and computational considerations in applying unsupervised methods to fraud detection.
Anshula Sharma, Duy H. Ho· 2026 6th International Confe...· 0 citations