This work is inspired by the paper Gated CNN and CNN stock price prediction
In this work I applied the Gated CNN on S&P500 stock price prediction. It beats several methods even without hyperparams tuning:
- Logistic Regression
- KNN
- SVM
- Random Forest
This method using CNN in a similar way as LSTM but with less training time needed.
Run the main.py program will give you the results. The layers specs is also listed in the main.py. Please feel free to play with those hyperparams.