Discussion of article "Gradient Boosting (CatBoost) in the development of trading systems. A naive approach" - page 9

[[Deleted]]

Roman Poshtar #:

No errors. MA 200

Code

Here is the log

And try pip update pandas

I'm using miniconda python 9.1, pandas is there, I'll post the version later

[[Deleted]]

Maxim Dmitrievsky #:
Try pip update pandas

I'm using miniconda python 9.1, pandas is there, I'll post the version later

>>> import pandas as pd

>>> pd.__version__

'2.2.0'

[Roman Poshtar]

Maxim Dmitrievsky #:
miniconda

I will try miniconda. THANK YOU!

[Roman Poshtar]

I think I know what the problem is. I don't have the machka values at large periods. I'll keep thinking. Thanks for your help.

[[Deleted]]

Roman Poshtar machka values at large periods. I'll keep thinking. Thanks for your help.

So there is not enough history, you need a bigger file. The size of the dataframe is always reduced by the period of the largest mashka, because the first values are not defined. But there is dropna(), which removes all NaNs

Strange why you are getting those columns returned.

[Roman Poshtar]

Maxim Dmitrievsky #:
So there is not enough history, you need a bigger file. The size of the dataframe is always reduced by the period of the largest Mashka, because the first values are undefined. But there is dropna(), which removes all NaN

Strange why you are getting those columns returned.

Maxim, can you tell me how to remove these columns from the final array. I am not good at Python. I want to try my own feature sets. Thank you.

[[Deleted]]

Roman Poshtar #:

Maxim, please tell me how to remove these columns from a finite array. I am not good at Python. I want to try my own feature sets. Thanks.

The get_prices function already does this at the end:

return prices.dropna()

[Roman Poshtar]

Maxim Dmitrievsky #:

The get_prices function already does this at the end:

I see, thanks.

[Roman Poshtar]

prices.dropna()

It didn't work at the end. The archive still contained Nan values. It was solved by simply deleting lines.

[Busingye Tusasirwe]

I seem unable to reproduce the results from the python tester. The MT5 tester is not reproducing the results for the same period in python tester.

Otherwise, I ported the model as explained. 

I put cat_model.mqh and cat_trader.mql5(compiled to .ex5).

But results look different.