Work in progress...
The goal of this project is to use reinforcement learning to optimize the portfolio which consists of stocks and cryptocurrencies. The portfolio is optimized by using a neural network to predict the future value of the portfolio.
Stock data is obtained from Yahoo Finance. We have collected stock data for S&P 500 companies. We have created a script to download the data from Yahoo Finance. The data for each company is stored in a separate CSV file and we have merged all the CSV files into one CSV file. The data is stored in the utils/datasets/all_stocks_5yr.csv file. we have used the data from the utils/datasets/all_stocks_5yr.csv file.
Cryptocurrency data is obtained from Yahoo Finance. We have collected top 9 cryptocurrencies on the basis of market capitalization. The data is obtained from the Yahoo Finance API. We have used the same script to download the data from Yahoo Finance. The data for each cryptocurrency is stored in a separate CSV file and we have merged all the CSV files into one CSV file. The data is stored in the utils/datasets/all_crypto_5yr.csv file. We have used the data from the utils/datasets/all_crypto_5yr.csv file.
You can use the script to download last n years of data for a list of stocks or cryptocurrencies from Yahoo Finance. Follow the steps mentioned here
We considering two model architectures for stock & cryptocurrencies price prediction. The first model is a LSTM based RNN model. The second model is a CNN.
Clone this repository using the following command:
git clone https://github.com/PacificG/Portfolio-Optimization-using-stocks-and-cryptocurrencies.gitIn the Portfolio-Optimization-using-stocks-and-cryptocurrencies directory, run the following command to set up the environment:
virtualenv -p python3 venvActivate the virtual environment using the following command:
source venv/bin/activate # for linux
venv\Scripts\activate # for windowsInstall the required packages using the following command:
pip3 install -r requirements.txt