TrendMaster is an advanced stock price prediction library that leverages Transformer deep learning architecture to deliver highly accurate predictions, empowering investors with data-driven insights.
- Features
- Why TrendMaster?
- Installation
- Quick Start
- Sample Results
- User Interface
- Documentation
- Contributing
- License
- Show Your Support
- Contact
- More from HJ Labs
- Try Our Algo Trading Platform hjAlgos
- Advanced Transformer-based prediction model
- High accuracy with mean average error of just a few percentage points
- Real-time data visualization
- User-friendly interface
- Customizable model parameters
- Support for multiple stock symbols
TrendMaster stands out as a top-tier tool for financial forecasting by:
- Utilizing a wealth of historical stock data
- Employing sophisticated deep learning algorithms
- Identifying patterns and trends beyond human perception
- Providing actionable insights for smarter investment strategies
Get started with TrendMaster in just one command:
pip install TrendMasterHere's how to integrate TrendMaster into your Python projects:
# Example usage of merged_module.py
from trendmaster import (
DataLoader,
TransAm,
Trainer,
Inferencer,
set_seed,
plot_results,
plot_predictions
)
import pyotp
# Set seed for reproducibility
set_seed(42)
user_id = 'YOUR_ZERODHA_USER_ID'
password = 'YOUR_ZERODHA_PASSWORD' # Replace with your password
totp_key = 'YOUR_ZERODHA_2FA_KEY' # Replace with your TOTP secret key
# Generate the TOTP code for two-factor authentication
totp = pyotp.TOTP(totp_key)
twofa = totp.now()
# Initialize DataLoader and authenticate
data_loader = DataLoader()
kite = data_loader.authenticate(user_id=user_id, password=password, twofa=twofa)
# Prepare data
train_data, test_data = data_loader.prepare_data(
symbol='RELIANCE',
from_date='2023-01-01',
to_date='2023-02-27',
input_window=30,
output_window=10,
train_test_split=0.8
)
import torch
# Initialize model, trainer, and train the model
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
print(f'Training on {device} device.')
model = TransAm(num_layers=2, dropout=0.2).to(device)
trainer = Trainer(model, device, learning_rate=0.001)
train_losses, val_losses = trainer.train(train_data, test_data, epochs=2, batch_size=64)
# Save the trained model
trainer.save_model('transam_model.pth')
# Initialize inferencer and make predictions
inferencer = Inferencer(model, device, data_loader)
predictions = inferencer.predict(
symbol='RELIANCE',
from_date='2023-02-27',
to_date='2023-12-31',
input_window=30,
future_steps=10
)
# Evaluate the model
test_loss = inferencer.evaluate(test_data, batch_size=32)Evaluate the performance of TrendMaster using our comprehensive backtesting framework. Our Transformer-based model has been rigorously tested to ensure reliability and accuracy in diverse market conditions.
Explore detailed backtest results on our hjAlgos Backtest Platform.
Sample Backtest Performance Chart
Our Transformer-based prediction model demonstrates impressive accuracy:
TrendMaster comes with a sleek, user-friendly interface for easy data visualization and analysis:
For detailed documentation, including API reference and advanced usage, please visit our Wiki.
We welcome contributions! Please see our Contributing Guidelines for more details.
This project is licensed under the MIT License - see the LICENSE file for details.
If you find TrendMaster helpful, please consider giving it a star on GitHub. It helps others discover the project and motivates us to keep improving!
For questions, suggestions, or collaboration opportunities, please reach out:
- Website: hjlabs.in
- Email: [email protected]
- LinkedIn: Hemang Joshi
Check out our other exciting projects:
Ready to elevate your trading strategy?
Created with ❤️ by Hemang Joshi