Machine learning in trading: theory, models, practice and algo-trading - page 664

 
elibrarius:
You can also do it in binary form to save money. But I don't have time to develop a replacement for what I have. Other ideas must be tested... It will be more like tuning) And it will be necessary, if something on NS will vouch for earnings.

And what is there to save if we get 1.5 GBytes/s with all conversions taken into account? ) And everything is very simple and uniform.

elibrarius:
So far it has been necessary to output to the terminal, with crutches, but I have done it. I should do something more intelligent - may be I will also do in R, I'm still a beginner in it and do not know all the advantages.

By the way, there are some good books on the basics of R on the web:

1. Robert R. Kabakov R in action. R data analysis and visualization.

2. R in a Nutshell by Joseph Adler (in English, but you can understand it, if necessary, at least in parts)
 
Yuriy Asaulenko:

And what is there to save if we get 1.5 GBytes/s with all conversions taken into account? ) And everything is very simple and uniform.

By the way, there are some good books on the basics of R on the web:

1. Robert R. Kabakov R in action. Data analysis and visualization in R.

2. R in a Nutshell by Joseph Adler (in English, but you can understand it if necessary, at least in parts)

here is more

https://msperlin.github.io/pafdR/

But you don't need it all, you may code everything in MT5 and become an octopus

Processing and Analyzing Financial Data with R
  • Marcelo S. Perlin (marcelo.perlin@ufrgs.br)
  • msperlin.github.io
Welcome to the book site of Processing and Analyzing Financial Data with R. Here you can find the full online content of the book. The book is also available as an ebook (Amazon) and hardcover (CreateSpace / Amazon). If you liked the material and are using the book, please consider purchasing it. As an author, I certainly appreciate the gesture...
 
Yuriy Asaulenko:

And what is there to save if we get 1.5 GBytes/s with all conversions taken into account? ) And everything is very simple and uniform.

By the way, there are some good books on the basics of R on the web:

1. Robert R. Kabakov R in action. Data Analysis and Visualization in R.

R in a Nutshell by Joseph Adler (in English, but you can understand it, at least in parts, if necessary).
Thanks! I've read the basics, and the rest will pick up as needed.
 
SanSanych Fomenko:

This is not the first time you've written about files.

And how is the problem of readiness polling solved?

Yes, I preached this idea even before the appearance of R-MT-DLL.) In terms of speed no problems, in terms of convenience of modification at the stage of development surpasses DLL.


Semaphore files. So, we write info file. When ready we write a semaphore file. Appearance of the semaphore file indicates that the data is ready for reading.

If we read the info file then we remove it and remove the semaphore file. The disappearance of the semaphore file signals that the receiver is ready to receive the next dose of information.

The files are of course different.

Several different semaphore systems can be used, e.g. for duplex channels, question-answer type.

Yes, the topic is about RAM-Disk There is also a software for testing.

My RAM-Disk has only 64K of memory, i.e. nothing at all).

 
Maxim Dmitrievsky:

here's more

https://msperlin.github.io/pafdR/

The literature ishere.

The search works fine for a wide variety of keywords.

Books are paid, usually 10 rubles, so you buy unlike torrents.

R
R
  • www.twirpx.com
Библиотека. Компьютерная литература. R
 
SanSanych Fomenko:

The literature should be drawnfrom

The search works fine for a wide variety of keywords.

Books are paid, usually 10 rubles, so you buy unlike torrents.

No, I'm done for the spring ) here are a couple of bots lazy to finish writing to March and all

I have nothing more to lose in this subject, especially to dig around in R. If the greatest minds and programmers break their spears, I have nothing to do there

The theory is more or less understandable. And just to use some libs without understanding the source code won't lead to anything good.

or you'll end up like this https://www.mql5.com/ru/forum/217284/page70#comment_6484707

Интересное и юмор (политика и история под запретом)
Интересное и юмор (политика и история под запретом)
  • 2018.02.01
  • www.mql5.com
При публикации, пожалуйста, соблюдайте правила ресурса...
 
Maxim Dmitrievsky:

Nah, I'm done for the spring ) just a couple of bots I'm lazy to finish by March and that's it

I have nothing more to lose in this subject, especially if I'm going to dig around in R. If the greatest minds and programmers break their spears, I have nothing to do there

The theory is more or less understandable. And just to use some libs without understanding the source code won't lead to anything good.

otherwise you'll end up like this https://www.mql5.com/ru/forum/217284/page70#comment_6484707

What's in the spring? A vegetable garden, potatoes? )))
 
Yuriy Asaulenko:

I don't understand this. Why do you need to transfer such volumes to MT from R? And what to do with them there?

I have a different concept. From MT only necessary market information, in MT - information for orders and deals. The rest is implemented in Jave, C++, Python, R, PHP, etc.

I don't understand why you have to make a zoo. MT terminal - let it solve the terminal tasks - market data and orders-transactions.

Absolutely correct approach.

 
elibrarius:
For example, to display all predictions of the NS on the training area and forward on the chart and visually assess whether I want from the NS or it is trained bullshit, which accidentally turned out to be winning at a particular section of history.

R has shiny for that. Output any kind of graphics.

 
elibrarius:
What's in the spring? Garden, potatoes? )))

No, girls sea sun... maybe Australia and New Zealand finally, have long wanted to go there :)

If there will be a cool TC on the NS and need investment, let me know :))