Research Article
Fraud Detection in Financial Transactions: A Comparative Study of Machine Learning Models with Ensemble Voting
@INPROCEEDINGS{10.4108/eai.28-4-2025.2358048, author={Boddu Krishna Chaitanya Sravanthi and Sai Tejaswini Keerthi and Vempada Latha and Guddati Vijaya Lakshmi and Ch.V.V. Satyanarayana and Kosanam Vennela}, title={Fraud Detection in Financial Transactions: A Comparative Study of Machine Learning Models with Ensemble Voting}, proceedings={Proceedings of the 4th International Conference on Information Technology, Civil Innovation, Science, and Management, ICITSM 2025, 28-29 April 2025, Tiruchengode, Tamil Nadu, India, Part II}, publisher={EAI}, proceedings_a={ICITSM PART II}, year={2025}, month={10}, keywords={real-time fraud detection machine learning financial transactions synthetic minority over-sampling technique (smote) ensemble learning random forest gradient boosting support vector machine (svm)}, doi={10.4108/eai.28-4-2025.2358048} }- Boddu Krishna Chaitanya Sravanthi
Sai Tejaswini Keerthi
Vempada Latha
Guddati Vijaya Lakshmi
Ch.V.V. Satyanarayana
Kosanam Vennela
Year: 2025
Fraud Detection in Financial Transactions: A Comparative Study of Machine Learning Models with Ensemble Voting
ICITSM PART II
EAI
DOI: 10.4108/eai.28-4-2025.2358048
Abstract
Real-time fraud detection in financial transactions presents great importance for the stability of financial systems. The investigators proposed a machine-learningbased model to classify fraudulent activities in financial transactions using a dataset with 6.3 million rows. The methodology involves extensive data preparation, including removal of irrelevant columns, encoding of categorical variables, scaling of the data, and adaptation of the imbalances presented within the response classes using the Synthetic Minority Over-sampling Technique (SMOTE). Various machine-learning models, such as Random Forest, Gradient Boosting, Logistic Regression, and Support Vector Machine (SVM), are used to detect fraud. A voting ensemble is used to enhance the predictive quality and robustness. After model evaluation on proper metrics outlines such as the model-comparison precision and appropriateness, this verifies a good realtime fraud detection system. Results to date suggest that the system has increased the detection rate of fraudulent transactions significantly, making it an impressive contender for real-time fraud examination within the context of finance.
