Fraud Prediction and Verification of Smart Credit Card using Machine Learning Techniques

Abstract

Listen

Translate article

This study unveils a powerful method for smart credit card fraud detection and verification. This system integrates data preprocessing, feature engineering, and real-time prediction using a hybrid model that incorporates supervised machine learning algorithms, an encoder, and LSTM networks. A supervised LSTM network sorts transactions, while an unsupervised Autoencoder finds outliers. Assessment criteria strike a balance between recall and accuracy. Alerts are sent by the system upon detection of fraud, and it runs in real-time. Compliance, scalability, and constant monitoring are key points. To close the gap between ease and safety in contemporary monetary transactions, this project offers a state-of-the-art method for strengthening the security of smart credit cards.