A neural network-powered trading bot that uses LSTM (Long Short-Term Memory) to predict cryptocurrency prices and backtest a simple buy/sell strategy based on those predictions.
- Fetches historical price data from Yahoo Finance (or generates synthetic data)
- Preprocesses and scales time series data
- Builds and trains a deep learning model using LSTM layers
- Predicts future prices
- Generates buy/sell signals
- Simulates trades and backtests strategy
- Visualizes predictions, signals, and portfolio performance
- Python
- TensorFlow / Keras
- Scikit-learn
- Pandas
- NumPy
- Matplotlib
- Yahoo Finance (
yfinance)
git clone https://github.com/yourusername/neural-network-trading-bot.git
cd neural-network-trading-bot
pip install -r requirements.txt
⚙️ Usage
You can run the bot directly using:
bash
Copy
Edit
python trading_bot.py
To customize:
python
Copy
Edit
bot = NeuralNetworkTradingBot(
symbol='BTC-USD',
start_date='2020-01-01',
use_synthetic_data=True # Set to False to fetch real data
)
bot.run()
📊 Backtesting Metrics
The bot calculates and displays:
Total return
Annualized return
Sharpe ratio
Buy/Sell actions
Portfolio value over time
📁 Project Structure
bash
Copy
Edit
.
├── trading_bot.py # Main bot logic
├── README.md # This file
├── requirements.txt # Python dependencies
✅ Requirements
numpy
pandas
matplotlib
scikit-learn
yfinance
tensorflow
You can install them with:
bash
Copy
Edit
pip install -r requirements.txt
📈 Strategy Logic
Buy Signal: If predicted price is >1% higher than previous close.
Sell Signal: If predicted price is >1% lower than previous close.
📌 Limitations
The model is for educational purposes and may not perform well in real markets.
No risk management or transaction fee modeling.
Real-world trading requires more advanced strategy design and evaluation.
📄 License
This project is licensed under the MIT License.
👤 Author
Your Name – Sadman Shoumik Rouf
Disclaimer: This software is for educational purposes only and is not financial advice.