📈 AI Multi-Model Trading Bot

A cryptocurrency trading signal prediction system using an ensemble of 8 ML/DL models with real-time Streamlit dashboard.

---

🎯 Features

• 8 ML/DL Models: Logistic Regression, Random Forest, XGBoost, LSTM, GRU, CNN, LSTM+Attention, Transformer
• Ensemble Voting: Combines all models for robust predictions
• SHAP Explainability: Understand which features drive predictions
• Real-time Dashboard: Live price data and trading signals
• Hugging Face Integration: Models hosted on HF Hub

---

🖥️ Demo

Feature	Screenshot
Live Predictions	Real-time BUY/SELL/HOLD signals
Model Agreement	Visual gauge showing vote consensus
SHAP Analysis	Feature importance visualization
Price Chart	Interactive candlestick chart

---

🚀 Quick Start

1. Clone the Repository

git clone https://github.com/YOUR_USERNAME/Multi-Model-AI-Trading.git
cd Multi-Model-AI-Trading

2. Install Dependencies

pip install -r requirements.txt

3. Run the Dashboard

streamlit run app.py

The app will automatically download models from Hugging Face on first run.

---

🤖 Models

Type	Models
Traditional ML	Logistic Regression, Random Forest, XGBoost
Deep Learning	LSTM, GRU, CNN, LSTM+Attention, Transformer
Ensemble	Majority voting across all models

Models are trained on BTC-USD daily data from 2015-2025.

---

📊 Technical Indicators

The models use 10 features:

Indicator	Type
RSI	Momentum
MACD & Signal	Trend
Bollinger Band Width	Volatility
ATR	Volatility
Distance from SMA50	Trend
OBV % Change	Volume
ADX	Trend
Stochastic RSI (K & D)	Momentum

---

🔗 Hugging Face

• Models: AdityaaXD/Multi-Model-AI-Trading-Bot
• Dataset: AdityaaXD/Multi-Model-Trading-Data

---

📁 Project Structure

├── app.py                    # Streamlit dashboard
├── multi_model_trading_system.ipynb  # Training notebook
├── requirements.txt          # Dependencies
├── PROJECT_REPORT.md         # Detailed report
├── upload_to_huggingface.py  # HF upload script
└── upload_data.py            # Dataset upload script

---

🛠️ Training Your Own Models

# Run the training notebook
jupyter nbconvert --execute multi_model_trading_system.ipynb

# Upload to Hugging Face (optional)
python upload_to_huggingface.py

---

📈 Performance

Metric	Best Model
Accuracy	Random Forest (52.2%)
Sharpe Ratio	Logistic Regression (1.94)
Total Return	Transformer (+353%)

---

⚠️ Disclaimer

This project is for educational and research purposes only. It should NOT be used for actual trading decisions. Cryptocurrency markets are highly volatile and past performance does not guarantee future results.

---

🤝 Contributing

Contributions are welcome! Feel free to open issues or submit pull requests.

---