Pnny13/fraud-detection-model

🔥 Death Legion Fraud Detection System

🎯 Overview

Welcome to the Death Legion Fraud Detection System - a state-of-the-art machine learning solution for real-time credit card fraud detection. Built by the Best Teams with cutting-edge Random Forest technology, this model achieves exceptional performance on highly imbalanced financial datasets.

🚀 Deploy This Space

🎮 Use This Model

⚡ Key Features

• 🛡️ Advanced Random Forest Architecture: 500 estimators with optimized depth
• 📊 Superior Performance: AUPRC 0.8177 on imbalanced fraud data
• 🔒 Secure Safetensors Format: Fast, safe model serialization
• ⚡ Real-time Inference: Sub-millisecond prediction latency
• 🎯 Imbalanced Data Optimized: Precision-Recall focused evaluation

📈 Performance Metrics

Metric	Score	Status
AUPRC	0.8177	✅ Excellent
Precision	0.8182	✅ High
Recall	0.8265	✅ Strong
F1-Score	0.8223	✅ Balanced

🏗️ Model Architecture

Random Forest Classifier
├── Estimators: 500
├── Max Depth: 25
├── Features: 30 (V1-V28 + Time + Amount)
├── Classes: 2 (Legitimate, Fraudulent)
└── Format: Safetensors (6.12 MB)

📖 How to Use

Quick Start

1. Access the Live Demo: Visit https://huggingface.co/spaces/Pnny13/fraud-detection-space

2. Enter Transaction Features: Input the 30 features (V1-V28, Time, Amount) from your credit card transaction

3. Get Instant Prediction: The system will return:

   • Fraud probability score (0-100%)
   • Binary classification (Fraud/Legitimate)
   • Recommendation for action

Programmatic Usage

from safetensors.numpy import load_file
import numpy as np

# Load model from Hugging Face Hub
from huggingface_hub import hf_hub_download
model_path = hf_hub_download(
    repo_id="Pnny13/fraud-detection-model",
    filename="fraud_detector.safetensors"
)

# Load and predict
tensors = load_file(model_path)
# ... use model for predictions

📁 Repository Structure

Pnny13/fraud-detection-space/
├── app.py              # Gradio application
├── requirements.txt    # Python dependencies
└── README.md           # Documentation

Pnny13/fraud-detection-model/
├── fraud_detector.safetensors  # Trained model
├── scaler.joblib               # Feature scaler
├── inference.py                # Prediction script
└── README.md                   # Model documentation

🔬 Technical Details

Dataset

• Source: Kaggle - Credit Card Fraud Detection
• Transactions: 284,807 total
• Fraud Cases: 492 (0.172% - highly imbalanced)
• Features: 30 PCA-transformed features + Time + Amount

Preprocessing

• Robust Scaling: Applied to Time and Amount features
• Feature Engineering: PCA components V1-V28
• Class Balancing: Optimized for precision-recall tradeoff

Training Configuration

RandomForestClassifier(
    n_estimators=500,
    max_depth=25,
    class_weight='balanced_subsample',
    random_state=42,
    n_jobs=-1
)

🎮 Live Demo

Try the live interactive demo at: https://huggingface.co/spaces/Pnny13/fraud-detection-space

🤝 Credits

Powered by Death Legion
Elite Machine Learning Division

Best Teams Collaboration
Excellence in AI Engineering

📜 License

This model is released under the MIT License. Use responsibly for fraud detection and financial security applications.

🔗 Links

• 🤗 Model: https://huggingface.co/Pnny13/fraud-detection-model
• 🚀 Space: https://huggingface.co/spaces/Pnny13/fraud-detection-space
• 📊 Dataset: https://www.kaggle.com/datasets/mlg-ulb/creditcardfraud

_{Built with 🔥 by Death Legion | Best Teams Elite Division}