Summer internship 2019: Fraud detection in value based cards

  • Prasad, Akshar (CoI)

Project Details


Credit cards, while convenient and attractive, are susceptible to frauds, leading to data and privacy breaches. Identifying fraudulent transactions in a cost effective and timely manner is a challenge for companies supplying these cards. Machine learning models provide an easy method to quickly identify anomalies in credit card transactions. In this paper, three models, Random Forest, Long Short Term Memory and Convolutional Neural Networks on High-Performance Computing Cluster (HPCC) distributed Systems computing platform have been used to classify a series of transactions as ‘Fraudulent’ or ‘Non fraudulent’. Although all three models give outstanding results, Random Forests and Convolutional Neural Networks perform the best, with 0.99 precision, recall and F1 score
Effective start/end date06/1/1909/1/19


Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.