Machine learning for robots - page 3

 
Evgeny Raspaev:
I've read the forum thread 2 times and still don't understand what you need. Please send me the source code, send me a description of the model, give me some room for experiments. I was training a neural network using MACD signals but I didn't see an impressive model. It would be interesting to see the model from your side.

All I need is to test a new technology, for that I am creating Expert Advisors for everyone who wants to use the templates drawn by them, draw them I will do for you too, and if you want to do it yourself, study the models on a fundamental level, here is a link to an article, there is a block diagram, libraries and examples on python.

 
Ivan Negreshniy:

All that I need is to test a new technology, for this I create Expert Advisors, all who wish to draw on templates drawn by them, draw will do and you, but if you want to engage in the MO, study the model at the fundamental level, here's a link to the article, there block scheme, libraries and examples on python.

What's in it for us? Try it out. If you want a collective discussion, make a neural network on the same ALGLIB, people will come up and there will be a discussion

All you're offering is the creation of a black box, the most you can get from your "help" ... Well, a pretty picture in the tester, programmers do not need it (90% of old-timers on the forum have such pictures in the tester), and those people who do not understand the whole process may naively suffer

 
Igor Makanu:

What's in it for us? Try it out. If you want a collective discussion, make a neural network on the same ALGLIB, people will come up and there will be a discussion

All you're offering is the creation of a black box, the most you can get from your "help" ... Well a pretty picture in the tester, programmers do not need it (90% of old-timers on the forum have such pictures in the tester), and those people who do not understand the whole process may naively suffer

ALGLIB is only one of the libraries, and if you know only it, then naturally, the rest will be black box to you, and if you understand it, then even my analogue of alglib, random forest with trees, fully converted to MQL operators, will be more understandable for review and analysis and faster in execution, and I would rather not talk about alglib neural network, in any way.
 
Ivan Negreshniy:

All I need is to test new technology, for that I create EAs for everyone who wants to use templates drawn by them, draw them will do it for you too, and if you want to do it yourself, study models on fundamental level, here is a link to an article, there is a block diagram, libraries and examples on python.

Curious architecture. Can't get it for yourself anywhere - to try it out? not for com. use

Ahhhh... you're co-authoring... cool )

 
Maxim Dmitrievsky:

Curious architecture. You can't get it for yourself anywhere - to try it out? not for com. use

Ahhhh... you're co-authors... cool )

There are separate libraries purely for Python x32 or x64, as well as a free engine with built-in Python API and linking on NamedPipes with MQL - https://www.mql5.com/ru/forum/261479/page16#comment_8011085, the only thing, as I wrote earlier - NDA, does not yet allow to publish it all on GitHub.
Собираю команду для развития МО (Дерева решения/леса) применительно к трендовым стратегиям
Собираю команду для развития МО (Дерева решения/леса) применительно к трендовым стратегиям
  • 2018.07.07
  • www.mql5.com
Предлагаю сплотиться для решения задачи МО применительно к трендам, т.е...
 
Ivan Negreshniy:
There are separate libraries purely for Python x32 or x64, as well as a free engine with built-in Python API and linking on NamedPipes with MQL - https://www.mql5.com/ru/forum/261479/page16#comment_8011085, the only thing, as I wrote earlier - NDA, does not allow to publish it all on GitHub yet.

i.e. i need to sign something with you and it will be possible to use it on python without any layers? if it is really cooler than the same forest, why not buy it (and from the screenshots you can see that it is cooler)

 
Maxim Dmitrievsky:

i.e. i need to sign something with you and i can use it on the python without any layers? if it's really cooler than the same forest, why not buy it (and from the screenshots you can see it's cooler)

all for free and without any subscriptions, just looking for somewhere, except GitHub and SourceForge, instalniki put, because here all the time call my links commercial)))
 
Ivan Negreshniy:
All free and without any subscriptions, just looking for somewhere, besides GitHub and SourceForge, to put the installers, because here all the time they call my links commercial))

googledisk then )

 
Maxim Dmitrievsky:

google disk then )

No problem for you, it's all right for private, we're talking about a public posting with documentation, upgrades, feedback for bug fixes and maintenance.
 
Ivan Negreshniy:
No problem for you, it can be private, we are talking about a public posting with documentation, upgrades, feedback for bug fixes and maintenance.

google code is still available, I don't know if it will work or not.

Reason: