Cryptocurrency Price Prediction and Backtesting

This repository contains two projects focused on cryptocurrency price prediction and backtesting: an LSTM-based User Interface which can be used to configure the LSTM model architecture & a more simplified Random Forest Regression Model. These projects utilise machine learning techniques to analyse historical cryptocurrency data and make predictions about future price movements.

LSTM Price Prediction Model

The LSTM (Long Short-Term Memory) price prediction model utilises deep learning to analyse historical price data and forecast future prices of a given cryptocurrency. The model is implemented using Python and TensorFlow, a popular deep learning library. The LSTM architecture is specifically designed to capture long-term dependencies in sequential data, making it well-suited for time series forecasting tasks.

The main features of the LSTM price prediction model include:

• Data collection: The model collects historical price data from the CoinGecko API, allowing it to analyse a wide range of cryptocurrencies.

• Data preprocessing: The collected data is preprocessed to remove missing values, normalise the values, and create input-output pairs for training the LSTM model.

• Model training: The LSTM model is trained on the preprocessed data, learning patterns and relationships in the historical price data.

• Model evaluation: The trained model is evaluated on a test dataset to assess its performance and accuracy in predicting future price movements.

• Price prediction: Once trained, the model can be used to make predictions on unseen data, providing insights into potential future price trends.

The LSTM price prediction model offers a powerful tool for cryptocurrency traders and enthusiasts to gain insights into price movements and make informed decisions.

Random Forest Regression Model

The Random Forest Regression script utilises the Random Forest algorithm to build a regression model for cryptocurrency price prediction. Random Forest is an ensemble learning method that combines multiple decision trees to make predictions. This script is implemented in Python using the scikit-learn library, a popular machine learning toolkit.

The main features of the Random Forest Regression model script include:

• Data collection: The script collects historical price and volume data from the CoinGecko API for a specific cryptocurrency.

• Feature engineering: The collected data is processed to create additional technical analysis features, such as moving averages and momentum indicators, which are known to be relevant for price prediction.

• Data preprocessing: The collected data is preprocessed by handling missing values, normalising the values, and splitting it into training and testing sets.

• Model training: The Random Forest Regression model is trained on the preprocessed data, learning patterns and relationships between the input features and the target variable (price).

• Model evaluation: The trained model is evaluated on the testing set to assess its performance and accuracy in predicting cryptocurrency prices.

• Residual analysis: The script analyses the residuals (differences between predicted and actual prices) to evaluate the model's performance and identify any systematic errors or anomalies.

The Random Forest Regression model provides a robust approach to cryptocurrency price prediction, leveraging a powerful ensemble learning technique to make accurate predictions.

Repository Structure

The repository is organised as follows:

• lstm - Directory containing all information related to the LSTM project

  • gif - Directory containing LSTM project gifs
  • img - Directory containing LSTM project images
  • lstm_analysis - Directory containing LSTM analysis
  • README.md

• resources - Directory containing 2 sub-directories

  • img - Directory containing all images used in this README
  • preparation - Directory containing all information related to the preparation of the project

• rfreg - Directory containing all information related to the Random Forest Regression project

  • img - Directory containing all images in the RFREG directory README
  • rfreg_analysis - Directory containing all information regarding the Random Forest Regression analysis
  • README.md

• src - Directory containing the 2 source code notebook files

  • LSTM_colab.ipynb - LSTM Google Colab Notebook
  • RFREG_colab.ipynb - RFREG Google Colab Notebook
  • README.md

Usage

To use the LSTM price prediction & backtetsing App or the Random Forest Regression Model:

• Clone this repository to your local machine or utilise the Google Colab shared file links above.

• Open the respective notebook using Jupyter Notebook, Google Colab or any compatible environment.

• Follow the instructions in the notebook to run the code cells and execute the models. Make sure to have an active internet connection to fetch data from the CoinGecko API.

• Analyse the results, including model performance metrics, plots, and predictions, to gain insights into cryptocurrency price movements.

• Experiment with different parameters, features, or models to further improve the performance and accuracy of the predictions.

• The provided notebooks are well-documented and guide you through the entire process, from data collection and preprocessing to model training and evaluation.

Dependencies

The projects in this repository require the following dependencies:

• Python 3.7
• TensorFlow
• scikit-learn
• pandas
• numpy
• matplotlib
• seaborn
• plotly
• ta
• ipywidgets

You can install the required dependencies using pip or any other package management tool.

Contribution

Contributions to this repository are welcome. If you have any suggestions, bug fixes, or additional features, feel free to open an issue or submit a pull request.

Happy cryptocurrency price prediction and backtesting!