Hi community members,
In this guided proyect for which there isn’t a given solution, I used many information of @ivelinagenova and @jafarinasim guided proyects. So firsly I want to thank them for sharing these great proyects. I invite you to visit their proyects: [here] (Predicting Stock Market Guided Project) and here respectively.
The two new contributions I have added in my proyect are:
A k-cross fold validation for the prediction model, that I realized it doesn’t have much sense in this proyect. I know it doesn’t help much but not all ideas are good. At least it could serve others not to use it. My conclusion about why this validation don’t work here, is because de dependency of the observations. But I would like to know your opinion too.
Another new idea that ocurred to me while working in this proyect was to try to make predictions further than one day ahead.
I think in the real world it would be very useful and profitable too.
This idea wasn’t easy to develop for me at all. If you are curious it is al the final section of my proyect and use a nested function. I know it is not easy to follow but I couldn’t find another easier way. Again, if you have another easier solution please tell me.
I hope my proyect would be useful for you if you are working in this proyect. Your feedback is welcome.
Many thanks, and sorry for any posible misspellings.