Problem Statement:

1. Real-Time Fraud Detection System

Tagline: “The Scam Stops Here – Be the Sherlock of Transactions”

Goal: Spot the wolves before they prey. Build a system so precise, it’s like

sniffing out the fraudsters’ next move in milliseconds.

Key Tasks:

§ Ingest live transaction data and run anomaly detection.

§ Implement machine learning (e.g., random forest, gradient boosting, deep learning) for continuous fraud scoring.

§ Showcase a user-friendly dashboard for real-time alerts.

Expectations:

§ Accuracy metrics (Precision, Recall, F1-Score).

§ Visualization of live fraud detection.

§ Documentation on model interpretability (e.g., SHAP, LIME).

References:

§ Kaggle Credit Card Fraud Detection Dataset

§ PCI DSS Guidelines

§ Stripe API Docs for Payment Data (example reference)