Machine Learning Strategies for Fraud Prevention in Financial Data

  • Unique Paper ID: 162615
  • Volume: 10
  • Issue: 10
  • PageNo: 540-545
  • Abstract:
  • The rapid expansion of the E-Commerce industry has led to an exponential surge in credit card usage for online transactions. Unfortunately, this growth has also resulted in an increase in fraudulent activities. Detecting fraud within credit card systems has become increasingly difficult for banks. Machine learning techniques play a pivotal role in identifying credit card fraud during transactions. To predict these fraudulent activities, banks employ various machine learning methodologies, leveraging historical data and incorporating new features to enhance predictive accuracy. In this study, we evaluate the effectiveness of three machine learning models—*Logistic Regression, **Decision Tree, and **Support Vector Machine (SVM)*—for credit card fraud detection. Our dataset comprises 3,925,159 credit card transactions sourced from Kaggle. Transactions are labeled as either "genuine" (denoted by "0") or "fraudulent" (denoted by "1"). With 3,921,920 genuine transactions and 3,239 fraud cases, the dataset is imbalanced. To address this, we create a new balanced dataset with 3,239 samples for training and testing the models. Our evaluation focuses on accuracy. The results indicate the following accuracy rates for the three models:- Logistic Regression: 92.47%, Decision Tree: 99.21%, SVM : 85.57% Comparatively, the Decision Tree outperforms both Logistic Regression and SVM. This research contributes to the ongoing efforts to combat credit card fraud using predictive analytics, artificial intelligence, and machine learning in real-time applications.

Cite This Article

  • ISSN: 2349-6002
  • Volume: 10
  • Issue: 10
  • PageNo: 540-545

Machine Learning Strategies for Fraud Prevention in Financial Data

Related Articles