The task of the project is to facilitate the work with the Stock-RNN project in order to predict the price of shares on the Polish stock exchange. The project is fully described in my blog code-addict.pl. The project consists of:
- Docker containers containing Jupyter Notebooks, Tensorflow (with Tensorboard), MongoDB
- Jupyter notebooks with code described in blog post.
- Project subtree - Stock-RNN which allow to build ,train and test neural network to predict Polish stock market prices.
To run project simply run docker container using:
docker-compose up
You will find http link to Jupyter Notebooks in the console.
There are two notebooks. They are both creating same Tensorflow graph and initialize training process but I thins PolishStockMarketPredictor_2
is better and I described it in my blog post.