Machine learning in trading: theory, models, practice and algo-trading - page 1371

New comment
 
Aleksey Vyazmikin:

What else is needed in Python besides "apply a model on this data and give an answer"? Optimization through automatic creation of parallel sockets?

to apply something, you first need to train something (c)

I don't know about optimization, i don't use MT optimizer at all, only bruteforce sometimes

connector is the easiest, the framework in 20 lines, the rest is written to fit your needs
 
Maxim Dmitrievsky:

to apply something, you need to train something first (c)

With optimization I don't know yet, I don't use MT optimizer at all, only bruteforce sometimes

I can teach it even not in python in my case, I can create model in different formats.

 
Maxim Dmitrievsky:


Connector is the simplest, the framework in 20 lines, the rest is filled in for your needs

Well, it would be interesting to see an example - you passed the value of a function as an array, and in response :)

 
Aleksey Vyazmikin:

Well, this will be interesting by example - you passed the value of the function as an array, and in response :)

Yes, that's how it is in the simplest case

50k characters are transferred and returned instantly. Calculation of some models depends purely on the model, there may or may not be delays

 
Aleksey Vyazmikin:

Then it is interesting to know the details of debugging such communication.

The mechanism is similar to the function call. The debugging is similar to the debugging of any program. I don't understand what you mean.
I use 2 sockets in parallel. In MKL one is enough - there everything is in one thread and the queue.
The task is to send/receive information quickly and release the terminal thread.
 
More worried about why all this is necessary and what models to saw now of all the great variety ))
 
Maxim Dmitrievsky:
I am more concerned about the question of why I need all this and what models to saw now of all the great variety.)
(Just use whichever one you like).
Imho, if you can not do simple, then you can not go to the complex).
 
Yuriy Asaulenko:
I'm not sure what I like, but I'm not sure what I like.)
Imho, if you can't do simple ones, then you don't have to go to complex ones).

There are simple ones in MT5 :)

 
Maxim Dmitrievsky:

there are simple ones in mt5 :)

I have no idea what is there). Since 08 I deal only with external to the terminal. With varying success).
 
Yuriy Asaulenko:
(Whatever suits your fancy, go for it.))
Imho, if you can not do simple, then you can not go to the complex).
I just saw Darch on the R. Found a couple of bugs, described in the comments. After a couple of weeks of silence this Darch ended up in the archive CRANa.
I asked the developer to fix and add some fixes, he did it. And then completely rolled back to the original version, erasing all the fixes. As a result, all modifications that I used were not available.
Conclusion - or do everything yourself, or use top products with very good support.
1...136413651366136713681369137013711372137313741375137613771378...3659
New comment