Machine learning in trading: theory, models, practice and algo-trading - page 759

 
Maxim Dmitrievsky:

google chrome has a button - translate the whole page into russian :) translates more or less well

))) sometimes I use it ) but I`m lazy in the first place )

 
Anatolii Zainchkovskii:

To start with at least a rough idea, divide the picture into 4 squares or 9 squares and evaluate. (a square like a pixel)

imho, too time consuming... these things take months to learn on teslas. 4 squares there is 0 information :)

 
Maxim Dmitrievsky:

imho, too time-consuming... these things take months to learn on teslas. 4 squares there is 0 info :)

That's why I started asking, I just have 9 squares and 7 lines, a simple check for a line to vadrate already leads to the number of pictures that the variables do not fit ...

 
Anatolii Zainchkovskii:

))) I sometimes use it) but laziness comes first)

https://github.com/yandexdataschool/Practical_RL/tree/master/

There are links to videos in Russian ... just if you can handle at least part of it, it will not be superfluous :)

 
Maxim Dmitrievsky:

https://github.com/yandexdataschool/Practical_RL/tree/master/

there are links to videos in russian... just if you can handle at least part of it, it won't be too much trouble :)

Thanks )

 
Anatolii Zainchkovskii:

That's why I started asking, I just have 9 squares and 7 lines, a simple check for a line to vadrate already leads to the number of pictures that the variables do not fit ...

Yes, and I need hundreds of thousands of pictures

 
Maxim Dmitrievsky:

with computer vision, it's a 100 layer grid.

there was a library used from a japanese site that supposedly used this grid in robot computer vision. (I could be wrong, don't worry) there was also an example of how to use the vision oval... The way it worked there only describes the libraries and the fact that the trader wanted to use them for pattern recognition. I'll send you the link when I get free. It's not so clear from my words.

 
Maxim Dmitrievsky:

Yes, and we need hundreds of thousands of pictures.

it's not a problem, i ran the history loop and got.... for 4 years of history on m15 you get 24000 pictures... Not enough for full set, of course, but I can fill 10 years of history...

 
Evgeny Raspaev:

there was a library used from a japanese website that supposedly used this grid in robot computer vision. (I could be wrong, don't worry) there was also an example of how to use the oval circle... The way it worked there only describes the libraries and the fact that the trader wanted to use them for pattern recognition. I'll send you the link when I get free. It's not so clear from my words.

As far as I know the most reasonable Open Source library is this one https://opencv.org/

OpenCV library
  • opencv.org
OpenCV (Open Source Computer Vision Library) is released under a BSD license and hence it’s free for both academic and commercial use. It has C++, Python and Java interfaces and supports Windows, Linux, Mac OS, iOS and Android. OpenCV was designed for computational efficiency and with a strong focus on real-time applications. Written in...
 
Anatoly Zainchkovskii:

no problem with it, I ran the history loop and got.... for 4 years of history on m15 it turns out 24000 pictures... It's not enough for a full set, of course, but you can do 10 years of history...

i couldn't find that article, and i found this neuron here

SOINN — самообучающийся алгоритм для роботов
SOINN — самообучающийся алгоритм для роботов
  • 2029.07.13
  • habrahabr.ru
Пост №1. Что такое SOINN SOINN – это самоорганизующаяся инкрементная нейронная сеть. Структура и алгоритм такой нейронной сети повидимому хорошо себя зарекомендовал в японской лаборатории Hasegawa (сайт — haselab.info), потому что он в итоге был взят за основу и дальнейшее развитие алгоритмов искусственного интеллекта шло путем небольших...
Files:
SOINN.zip  2850 kb