Discussion of article "Advanced resampling and selection of CatBoost models by brute-force method" - page 11

 
Maxim Dmitrievsky:

i.e. this is an example of loading data from the terminal and saving it to a file. And then you can use it in colab

Thank you.

There is one more question. How to open trades on this MO model directly in python? Can you send me the code?

[Deleted]  
Evgeni Gavrilovi:

Thank you.

There is one more question. How to open trades on this MO model directly in python? Can you send me the code?

There is no ready code, you need to write a class of opening orders for this model

You can read the help

https://www.mql5.com/en/docs/integration/python_metatrader5

Документация по MQL5: Интеграция / MetaTrader для Python
Документация по MQL5: Интеграция / MetaTrader для Python
  • www.mql5.com
Python является современным высокоуровневым языком программирования для разработки сценариев и приложений. Содержит множество библиотек для машинного обучения, автоматизации процессов, анализа и визуализации данных. Пакет MetaTrader для Python предназначен для...
 
Maxim Dmitrievsky

It doesn't work at Colab unfortunately(

https:// colab.research.google.com/drive/1jUasWYQ_Gf3R0AOQDwVJtkji_49QH_0y

The mt5 module is still specified there, and without get_prices() I don't know how to set start and stop_date. Where is the error there?

[Deleted]  
Evgeni Gavrilovi:

It doesn't work at Colab, unfortunately(

https:// colab.research.google.com/drive/1jUasWYQ_Gf3R0AOQDwVJtkji_49QH_0y

The mt5 module is still specified there, and without get_prices() I don't know how to set start and stop_date. Where is the error there?

You just open the file, get_prices is no longer needed. That is, the code after g_p remains. Well, for validation you also need to load your own file.

I don't understand why this is necessary, because colab is slower than even my laptop with a mobile processor.

 
Maxim Dmitrievsky:

just open the file, get_prices is no longer needed. I.e. the code after g_p remains. well, you also need to load your own file for validation.

I don't understand why this is necessary, because colab is slower than even my laptop with a mobile processor.

It all worked, 2 files cat_model.mqh are loaded in output, it's in /include/ and the second one - catmodel.h for which directory?

[Deleted]  
Evgeni Gavrilovi:

Everything worked out, 2 files cat_model.mqh are loaded in output, it is in /include/ and the second one - catmodel.h for which directory?

cat_model.mqh is an include file for the bot, it should be placed in the folder with terminal includers
 
Maxim Dmitrievsky:
cat_model.mqh is an include file for the bot, it should be thrown into the folder with terminal includers

Yes, I've uploaded it. but the tester doesn't start in metatrader, apparently because of the missing catmodel.h file.

Doesn't it take part in forming trading signals?

[Deleted]  
Evgeni Gavrilovi:

Yes, I've uploaded it. but the tester doesn't run in metatrader, apparently because of the missing catmodel.h file.

Doesn't it take part in the formation of trading signals?

No, it's a C library, .mqh is parsed from it.

 
Maxim Dmitrievsky:

No, it's a C library, it parses .mqh from it.

added file paths - trayning prnew.csv and test prnews.csv

but the received R2 is almost always higher than 0.9, maybe look_back is not set correctly and that's why the received mqh file is wrong, because of which I can't test in terminal.


https:// colab.research.google.com/drive/1eeyRA5bGaFMfX1THnMsL5hwKmxBkqvqP


https://drive.google.com/file/d/1LIRhpk5iU_dYQbefZ-FFQM6XMV_cOh26/view?usp=sharing test data


https://drive.google.com/file/d/18RpJec9EGSCSknwaHsevgHcZuCeoOvP5/view?usp=sharing training data

 
By the way, using get_prices in Jupyter Notebook, the mqh file loads correctly, then tests fine.