Stock Market prediction - The tabnet

Predicting stock value is really important to yield significant profit. Taking the features or factors that effect the yield, I have done a regression using the latest deep learning architecture ‘Tabnet’ and a xg boost, and used an ensemble of both the models for final prediction. The other problem is to predict put calls ratio which is a time dependent factor. I have used an RNN for predicting PCR of one day using past three days data. I tried a self supervised way to deal with the missing values!

wanna go deeper? Check my Github repo!!