I have been working on problems related to finance and machine learning. Until now I have used Gradient Boost Trees and LSTM algorithms for predictions. Are there other techniques used in the domain that give better results. Please do share your opinions regarding the techniques, feature engineering and the model architecture used. Thanks!
I have been working on this for quite some time and I have compared over 20 different type of models for stock price prediction from SVM, Deep Forest, LSTM etc. you wont get the best performance from these models especially when there is high volatility. The state of the art technique which outperforms them all is a Bidirectional Gradient Recurrent Unit (BGRU) followed by Generalized Additive Models.
Could you please link any results and performances of these models? Thanks!
I’m working on a formal release for Analyticsvidhya which will be ready by next week. If that’s too late for you I can email you some of the preliminary results with the python code.