Stock Price Prediction

#1

I’m getting negative accuracy.I’m tried to change various hyper parameters but not able to get reason for it.Point out my mistakes.
Link to jptr ntbk–
https://colab.research.google.com/drive/1RqKRbkl-xnSV-3HbH0YrPVOK6f2pbDXW

#2

Hey @Nik1, your code is correct, actually there are only 19 values, and that are also too close, so achieving accuracy in this competetion is very difficutl task until you model is completely perfect.

Recommended changes, use sklearn train_test_splti, add dropout layers, don’t use more than 2 lstm layers, 2 layers are sufficient for most tasks. Store the model, with lowest val_mae

Hope this resolved your doubt.
Plz mark the doubt as resolved in my doubts section. :blush:

#3

I hope I’ve cleared your doubt. I ask you to please rate your experience here
Your feedback is very important. It helps us improve our platform and hence provide you
the learning experience you deserve.

On the off chance, you still have some questions or not find the answers satisfactory, you may reopen
the doubt.

#4

I did the changes suggested by you but still getting negative accuarcy.Link to modified ntbk—https://colab.research.google.com/drive/1sv80LRky4KqVABxMPzmTxSfinbm3jSPe

#5

Hey @Nik1, yes now your file is completely fine, the only reason you model is getting less accuracy, is that when you calculate r2_score for such a less amount of points, it results in various error. We are trying to modify this challege to correct. Actually one easy hack have been applied by those higher on leaderboard, is that for the first 18 days, they intentionally put the next day value already given in x_text, and predicted only for those.

If your sole purpose is achieving accuracy, you can do this as well. Otherwise your file is fine, proceed with the course. I have asked the mentor to increase the number of points.

Hope this resolved your doubt.
Plz mark the doubt as resolved in my doubts section. :blush: