This is one of my CDIO projects (subjects) at my university.
The method is not effective for real-world application but it has the core idea.
That is learning the historical data and then try to predict the test set and the next day price.
I was not really writing each line of code but I tried to learn the meaning of pieces.
And I adjusted the code very much so that it could work better without changing variables too much.
There are still a lot of things to learn in AI, especially Deep Learning.