FraudShield: Detection of Fraud in Credit Card based on Machine Learning Techniques with integration of web-based Framework

FraudShield: Detection of Fraud in Credit Card based on Machine Learning Techniques with integration of web-based Framework

Tejas Prakash D, Grishma A V, Sneha M, Shweta A Nair, Sai Meghana J S, Dr. Kuppala Saritha

Dept. of IST

Presidency University

Banglore-64 India

ABSTRACT

In this paper, a general term for both fraud and theft committed with or while using a credit card for payment is "credit card fraud." In light of the daily increase in scammers, numerous different types of fraud are committed using credit cards. In order to address this issue, numerous methods like logistic regression, decision trees, KNN, and Naive Bayes algorithms are used. Various options are considered for this transaction, and the best one is implemented. Filtering out the above-mentioned tactics is done mostly to achieve the goal of detecting fraud and producing better results. Credit card fraud has increased due to the growing popularity of using credit cards for online purchases. The problem has been addressed through the use of machine learning algorithms to identify fraudulent transactions made with credit cards. In this paper, suggesting a system for detecting credit card fraud that employs machine learning techniques to distinguish between legitimate and fraudulent transactions. In order to examine previous transaction data and spot trends that point to fraudulent activity, the proposed system combines both supervised and unsupervised methods of learning. Additionally combining the developed machine learning methods with a web-based platform to deliver an intuitive user interface for immediate form of fraud detection. The test findings demonstrate that the suggested framework detects fraudulent transactions with a high degree of accuracy and with a minimal amount of false positives.

Keywords: Fraud Detection, Credit Card, Decision Tree, Logistic Regression, KNN, Naïve-Bayes, Streamlit, MLP, Accuracy, Precision, Macro average, Weighted average.