This project is part of my internship at ULiege on Deep RL in stock market trading with Professor Damien Ernst . Here I am validating the effectiveness of google trends data for an automated stock trading agent using the FinRL library.
If you want to change the ticker symbol and name for trends data, you can do it from the train_tune.py file by uncommenting the required timeframe and ticker symbol. Also, here is the link to my Medium article
Start with cloning the repository
git clone https://github.com/Athe-kunal/Reinforcement-learning-trading-agent-using-Google-trends-data.git
The jump to the directory
cd Reinforcement-learning-trading-agent-using-Google-trends-data/
Install all the dependences
pip install -r requirements.txt
Then if you want to download the pytrends data apart from what is present inside the Pytrends folder First do it to get help of CLI in downloading trends data
python pytrends_daily.py -h
Example for Apple to download data for month of October in 2021
python pytrends_daily.py -n 'Apple' -t 'AAPL' -say 2021 -sam 10 -soy 2021 -som 10 -c 0
Now you can run different cases and save results in your Account value folder. Also uncomment the dates that you want for testing in train_tune.py
python main.py -n 'Amazon' -t 'AMZN'
Note: The default environment is for a single stock trading only
SOME RESULTS
Sharpe ratio for different stocks
