XGBoost-Optimized Real-Time Financial Surveillance System for Fraudulent Activity Detection

Credit card fraud is important problem in today world [8], [14]. Fraudulent transactions in financial dealings must be done instantly. It must be identified to cut down losses. Conventional fraud detection systems generate a lot of false positives, which does not work and slow processing [12]. The Real-Time Credit Card Fraud Detection System developed here exploits Machine Learning [3], [7]. Data streaming in real-time through Kafka. Utilize cloud services for efficient detection of fraud transactions. Our solution combines XGBoost-based predictive modeling [1], [6]. We can make it better by processing the data as it comes in and alerting it automatically [11]. Having less delay in fraud detection accuracy. The system processes transaction records on the spot. Apache Kafka, evaluates them using an already-trained ML. Model is hosted on AWS EC2 and provides instant alerts. It improves its performance based on previous transactions detecting fraud [15]. Comparison through actual-world simulations show faster detection of frauds and minimizing false positives [2], [9]. The system further improves. Helps prevent fraud for financial security. Actively and easily used on a large scale. This enhances the use of AI in creating fraud-proof to find fraud in finance [10], [13]. Advanced programs were suitable to auto-adapt. This fraud is disabled using models and analytics that shift fraud strategies [4], [5].