TensorFlow-stocks-prediction-Machine-learning-RealTime
Predict operation stocks points (buy-sell) with past technical patterns, and powerful machine-learning libraries such as: Sklearn.RandomForest , Sklearn.Gradien
What is TensorFlow-stocks-prediction-Machine-learning-RealTime?
Predict operation stocks points (buy-sell) with past technical patterns, and powerful machine-learning libraries such as: Sklearn.RandomForest , Sklearn.GradientBoosting, XGBoost, Google TensorFlow and Google TensorFlow LSTM..Real time Twitter:
Overview of TensorFlow-stocks-prediction-Machine-learning-RealTime
What it is
This is an open-source project that predicts stock trading points (buy, sell, or do nothing) using past technical patterns and machine-learning libraries. It integrates models from Sklearn (RandomForest, GradientBoosting), XGBoost, Google TensorFlow, and TensorFlow LSTM. The project also supports real-time evaluation and alerts via Twitter, Telegram, and email.
Key features
- Tests 36 models with multiple feature combinations and model selections, expandable to TensorFlow, XGBoost, Sklearn, LSTM, GRU, dense, and linear models.
- Uses 637 technical stock indicators.
- Provides threshold and quality model evaluation.
- Selects the best technical patterns for each stock independently via neural networks.
- Uses a categorical target (buy, sell, do nothing) instead of a continuous target.
- Includes a real-time open-market evaluation system.
- Integrates with Twitter, Telegram, and Mail for alerts.
- Trains models with fresh daily stock data.
Use cases
- Stock trading signal generation for individual stocks.
- Backtesting and evaluating predictive models with historical data.
- Real-time monitoring and alerting for buy/sell opportunities.
- Research and experimentation with various machine-learning models for financial time series.
Getting started
The README emphasizes running the tutorial first. The tutorial file is located at Tutorial/RUN_buy_sell_Tutorial_3W_5min_RT.py. It guides through steps like reading OHLCV data, extracting technical patterns, calculating ground truth, normalizing data, creating time windows, splitting data, balancing classes, building a TensorFlow model, training, and evaluating. For detailed installation and usage, see the repository README.
License and pricing
The project is open-source. The author's license states that if you make a major improvement, you must notify the author via public or private contact ([email protected]), and the improvement will not be made public if you do not want it. The pricing type is listed as open_source. The specific license type is not stated in the metadata.
Based on the project's public README and repository metadata.