Problem Statement
E-commerce platforms are vulnerable to various forms of fraudulent activities, such as
unauthorized transactions, account takeovers, and fake reviews. Traditional rule-based
fraud detection systems may struggle to adapt to evolving fraud patterns and can
generate false positives, impacting the user experience.
Abstract
The E-commerce Fraud Detection using Machine Learning project aims to enhance the
security of online transactions by leveraging machine learning algorithms. Through the
analysis of user behavior, transaction patterns, and other relevant features, the system
seeks to identify and prevent fraudulent activities in real-time, providing a more adaptive
and accurate fraud detection mechanism.
Outcome
The outcome of this project is an advanced e-commerce fraud detection system
powered by machine learning. The system improves the accuracy of identifying
fraudulent transactions, reduces false positives, and enhances the overall security of
online shopping. E-commerce platforms benefit from increased trust among users, minimized financial losses due to fraud, and a more secure and reliable online shopping
experience.
Reference
Fraudsters find it easy to commit credit card fraud because it is an easy target. There has been an increase in online payment modes in due to e-commerce and other online platforms, there is now a higher danger of online fraud. Due to an increase in fraudulent online transactions, researchers have begun to evaluate and detect fraud using machine learning. In order to examine past transaction information and extract consumer behavioral patterns, our main goal in this study, a novel fraud detection algorithm for streaming transaction data is built and created. A system that clusters cardholders according to the amount of their transactions. In order to extract the behavioral pattern of the groups, we should aggregate the sliding window method transactions done by cards from various groupings. It is then decided which classifier with the best rating score can be chosen as one of the best methods to predict frauds after training different classifiers over the groups separately. The paper shows how the model is related to convolutional neural networks and afterward adding classifiers algorithms, for example, Isolation Forest, Local Outlier, and SVM can be utilized to recognize misrepresentation. As a result, concept drift can be solved via a feedback mechanism. We used the Kaggle credit card fraud dataset for this article.
- Apapan Pumsirirat and Yan Liu, “Credit card fraud detection using deep learning based on auto-encoder and restricted boltzmann machine,” International Journal of Advanced Computer Science and Applications, vol. 9, no. 1, 2018.
- Felipe Marino Moreno et al., “Transfer Learning. Solanki Jatin et al ‘A STEP FORWARD IN FRAUD DETECTION SYSTEM USING MACHINE LEARNING,” Journal of Critical Reviews, 2020. Google Scholar
- S. P. Devika, K. S. Nisarga, G. P. Rao, S. B. Chandini and N. Rajkumar, “A research on credit card fraudulent detection system,” International Journal of Recent Technology and Engineering, vol. 8, no. 2, pp. 5029-503, 2019.
- J. Huang and J. Liu, “Intrusion detection system based on improved BP neural network and decision tree,” 2012 IEEE 5th International Conference on Advanced Computational Intelligence, pp. 188-190, 2012.
- P. Raghavan and N. E. Gayar, “Fraud Detection using Machine Learning and Deep Learning,” 2019 International Conference on Computational Intelligence and Knowledge Economy (ICCIKE), pp. 334-339, 2020.
https://ieeexplore.ieee.org/document/10085493/