Skip to content

This repository contains an implementation of the LightGBM model for predicting financial instrument prices like stocks, currencies, and cryptocurrencies. It uses gradient boosting to analyze patterns in price data, aiming to enhance the accuracy and reliability of financial predictions.

License

Notifications You must be signed in to change notification settings

taleblou/LightGBM-Price-Prediction

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

6 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

LightGBM Model for Financial Predictions

This repository contains an implementation of an LightGBM model, specifically designed for predicting the prices of financial instruments such as currencies, stocks, and cryptocurrencies. The LightGBM algorithm leverages gradient boosting techniques, enabling it to capture intricate patterns in price movements and handle various dataset characteristics effectively. This approach enhances the accuracy and robustness of price forecasts across various datasets.

This is the original code sample for the LightGBM model. Explore my GitHub repository for additional models and implementations that cater to different financial prediction needs.

Performance Metrics

BTC-USD (Bitcoin)

Metric Open High Low Close
Mean Squared Error 0.003117 0.003003 0.003135 0.003120
Mean Absolute Error 0.032707 0.030720 0.032707 0.032671
R-squared 0.8679 0.8750 0.8653 0.8711
Median Absolute Error 0.015775 0.013368 0.014859 0.014888
Explained Variance Score 0.8866 0.8940 0.8853 0.8895

GC=F (Gold Futures)

Metric Open High Low Close
Mean Squared Error 0.002952 0.002961 0.002799 0.002867
Mean Absolute Error 0.042541 0.041599 0.041437 0.042491
R-squared 0.8526 0.8507 0.8605 0.8558
Median Absolute Error 0.035425 0.033177 0.034010 0.033128
Explained Variance Score 0.9053 0.9021 0.9126 0.9058

EURUSD (Euro/US Dollar)

Metric Open High Low Close
Mean Squared Error 0.000283 0.000226 0.000237 0.000277
Mean Absolute Error 0.012506 0.011682 0.011515 0.012449
R-squared 0.9377 0.9511 0.9486 0.9388
Median Absolute Error 0.010058 0.008580 0.010088 0.009810
Explained Variance Score 0.9379 0.9513 0.9487 0.9390

GSPC (S&P 500 Index)

Metric Open High Low Close
Mean Squared Error 0.001274 0.001255 0.001242 0.001345
Mean Absolute Error 0.030481 0.030138 0.029797 0.031667
R-squared 0.9066 0.9123 0.9082 0.9054
Median Absolute Error 0.030608 0.028235 0.028595 0.030568
Explained Variance Score 0.9525 0.9575 0.9510 0.9537

Related Websites

Free AI-powered short-term (5/10/30 days) and long-term (6 months/1/2 years) forecasts for cryptocurrencies, stocks, ETFs, currencies, indices, and mutual funds.

Get free trading signals generated by advanced AI models. Enhance your trading strategy with accurate, real-time market predictions powered by AI.

Discover free trading signals powered by expert technical analysis. Boost your forex, stock, and crypto trading strategy with real-time market insights.

About This Project

This LightGBM model is an initial implementation, released for public use. The project demonstrates the potential of deep learning models for financial predictions. While this repository focuses on LightGBM, I have also utilized other models, the code for which is available on my GitHub[https://github.com/taleblou/].

How to Use

  1. Clone this repository.
  2. Install the required libraries: pip install -r requirements.txt
  3. Prepare your dataset and follow the instructions in the notebook or script.
  4. Run the model and evaluate its performance using the provided metrics.

License

This project is open-source and available for public use under the MIT License. Contributions and feedback are welcome!

About

This repository contains an implementation of the LightGBM model for predicting financial instrument prices like stocks, currencies, and cryptocurrencies. It uses gradient boosting to analyze patterns in price data, aiming to enhance the accuracy and reliability of financial predictions.

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages