Abstract :

Online Financial Transactions Have Increased Significantly In Recent Years, Creating New Opportunities For Fraudsters And Posing Major Challenges To Digital Security. Traditional Fraud Detection Systems Often Struggle With Highly Imbalanced Datasets, Where Fraudulent Activities Represent Only A Small Fraction Of All Transactions. This Imbalance Reduces Model Accuracy And Increases False Negatives, Allowing Many Fraudulent Activities To Go Undetected. To Address This Issue, The Proposed System Introduces An Enhanced Online Fraud Detection Framework That Leverages Balanced Machine Learning Models Combined With Advanced Preprocessing Techniques.The System Utilizes Data Balancing Methods Such As SMOTE, Random Under-Sampling, And Hybrid Sampling To Create A More Representative Dataset For Training. Machine Learning Algorithms Such As Random Forest, XGBoost, Logistic Regression, And Support Vector Machines Are Then Employed To Classify Transactions Effectively. Feature Engineering And Correlation Analysis Are Applied To Improve Model Interpretability And Predictive Accuracy. The Framework Also Integrates Real-time Detection Capabilities, Allowing Suspicious Transactions To Be Flagged Instantly.Experimental Results Demonstrate That Balancing The Dataset Significantly Improves The Model’s Ability To Detect Fraudulent Activities While Reducing False Alarms. The Enhanced Models Show Superior Precision, Recall, And F1- Scores Compared To Traditional Approaches. Overall, This Research Presents A Robust And Scalable Fraud Detection System Capable Of Strengthening Online Transaction Security And Helping Financial Institutions Reduce Losses Due To Fraudulent Behavior. Keywords: Online Fraud Detection, Machine Learning, Imbalanced Data, SMOTE, Data Balancing Techniques, Classification Models, Real-Time Detection, Fraud Analytics, Predictive Modeling, Financial Security

Published:

18-12-2025

Issue:

Vol. 25 No. 12 (2025)

Page Nos:

338-347

Section:

Articles

License:

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

How to Cite