Skip to content

Latest commit

 

History

History
78 lines (64 loc) · 2.55 KB

README.md

File metadata and controls

78 lines (64 loc) · 2.55 KB

Fraud Detection System

Description

This project is a comprehensive fraud detection system designed to identify fraudulent transactions in a dataset. It combines a user-friendly dashboard for data visualization and analysis with a REST API for predictive modeling. Using machine learning techniques, the system identifies potential fraud and provides interactive insights.

Key Features

  • Streamlit Dashboard:
    • Visualize transaction data distribution.
    • Analyze fraudulent transactions by country and type.
    • Interactive world map visualization.
  • Machine Learning:
    • Random Forest Classifier for fraud detection.
    • Preprocessing with feature engineering.
  • REST API:
    • Predict fraud probability using Flask.

Technologies Used

  • Programming Language: Python
  • Libraries: Streamlit, Flask, Scikit-learn, Plotly, Pandas, NumPy
  • Visualization Tools: Plotly for interactive charts, Streamlit for dashboards
  • Model Deployment: Flask for API

Installation

Prerequisites

  • Python 3.8 or above
  • Required libraries (see requirements.txt)

Steps

  1. Clone the repository: 
    git clone https://github.com/your-repo/fraud-detection-system.git
  2. Navigate to the project directory: 
    cd fraud-detection-system
  3. Install the dependencies: 
    pip install -r requirements.txt
  4. Download the dataset and place it in the specified path:
    • Update the data_path variable in fraud_dashboard.py with the correct dataset location.

Usage

Streamlit Dashboard

  1. Run the Streamlit application: 
    streamlit run fraud_dashboard.py
  2. Open the URL provided by Streamlit in your browser (default: http://localhost:8501).
  3. Explore data visualizations, analyze fraud distribution, and view model results.

Flask API

  1. Start the Flask server: 
    python flask_api.py

Use the API to make predictions by sending a POST request to /predict.

Dataset

This project uses a transactional dataset that includes:

  • Features:
    • Time_step: Timestamp of the transaction.
    • USD_amount: Transaction amount in USD.
    • Sender_Account, Sender_Country: Sender details.
    • Bene_Country: Beneficiary country.
    • Transaction_Type: Type of transaction.
  • Target: Label (1 for fraud, 0 for non-fraud).

Future Improvements

  • Implementing advanced machine learning techniques like ensemble methods or deep learning.
  • Adding more robust error handling in the Flask API.
  • Including real-time transaction monitoring capabilities.