hsxavier/bovespa

Predicting (not) BOVESPA stock prices with an LSTM model

LSTM model for predicting stock prices

In this project I built a recurrent neural network model to predict future stock prices based on past prices. Spoiler alert: the results show that this is actually not possible.

Project structure

.
├── requirements.txt  <- Required python packages 
├── analises          <- Where all the code is saved.
└── dados             <- Folder where data resides.

Contents

All the code is stored in a jupyter notebook. Besides the tensorflow model building code, I also provide functions to download and process stock prices data from Brazilian stock market BOVESPA.

Author

• Henrique S. Xavier - @hsxavier

License

This project is distributed under MIT license - check the LICENSE file for more details.