*Result*: A NOVEL BEHAVIORAL BASED FRAUD DETECTION SYSTEM USING MACHINE LEARNING.

*Frauds are caused by increasing e-commerce platforms by developments of rapid commercial and technologies that effects harm of these platforms. Now a day’s credit cards usage is becoming highly popular, so, it is important to detect fraud to secure user accounts timely and accurately. To identify the frauds existing models are using manually process like original or aggregated features as their transactional representations but hidden behaviors of fraudulent are not identified. Because of fraudulent activity, company's reputation will be damaged and it leads to large financial losses, so, in financial industry fraud detection is becoming challenging. Sometimes, because of constantly developing methods that fraudsters use new frauds techniques, so to detect frauds conventional rule-based fraud detection systems will struggle so much. For online payment fraud detection, behavior-based techniques are recognized as a promising method. However, low-quality behavioral data is used for developing high resolution behavioral models and also it is a significant challenge. In financial transaction data the fraudulent patterns are identified because of it ability. Therefore, to overcome these limitations, Machine Learning (ML) algorithms are used and it is becoming popular because of its effectiveness. To determine fraudulent transaction is numbers of algorithms have been developed because of advanced machine learning and data science. A novel machine learning-based behavioral fraud detection system is introduced in this analysis and Support Vector Machine (SVM) algorithm is used to find the fraud. To evaluate this model's performance uses Accuracy, precision, recall, and F1-score parameters. [ABSTRACT FROM AUTHOR]*