FraudIQ – ML Risk Analysis Tool

FraudIQ is a machine learning-powered tool for assessing the risk of credit card transactions. Built as a full-stack prototype, it allows users to interact with a logistic regression model in real time and generate transaction risk reports.

🚀 Key Features

• Logistic Regression Model
  Trained on 284,807 credit card transactions with a validation accuracy of 97.5%.

• Risk Scoring System
  Calculates a fraud risk score (0–100) and classifies transactions into Low, Medium, or High Risk.

• Real-Time Analysis
  Users input transaction amount and time, and receive immediate risk evaluations.

• Risk Summary Report
  Generates a downloadable PDF summarizing the number of transactions, risk distribution, average score, and max score.

🧑‍💻 Tech Stack

• Frontend: React.js
  Interactive dashboard with chart visualizations using Recharts.

• Backend: Flask
  REST API serving real-time predictions and generating risk reports.

• Machine Learning:
  Logistic Regression using scikit-learn, with SMOTE to address class imbalance.

📊 How It Works

1. User Input
   Enter transaction amount and time through the React dashboard.

2. Prediction
   Backend model returns a risk score and level based on the trained logistic regression model.

3. Visualization
   View results through a dynamic chart and summary card.

4. Report Generation
   Instantly generate a downloadable PDF with all relevant transaction risk data.

🖥️ Installation & Setup

# Clone the repository
git clone https://github.com/your-username/FraudIQ.git
cd FraudIQ

# Install backend dependencies
pip install -r requirements.txt

# Run the backend
python app.py

# Navigate to frontend directory
cd client

# Install frontend dependencies
npm install

# Start the React frontend
npm start