Machine learning in trading: theory, models, practice and algo-trading - page 886
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
Can't you just flip through the book and use it as a reference? That's why I don't answer your questions.)
In the book you suggested Rattle is mentioned 2 times! So it did not help me, the actual information in Russian on this package could not find, so I asked here.
I may have downloaded the log, went to R, and there to divide the frame, I do not know how. I do not know how to divide the frame, so then what to do with the log, where to insert it?
Or here"Open R itself and there load an excel file - this is one line. Then divide the data frame into two."Again, I don't know how to divide it in R, so I divide it in excel. What do I do next?
You can't flip through the book and use it as a reference? So I did not answer your questions.)
There are all these examples in articles on NS here.
You are given a direction, and no one will write every line. I thought you said you've read it, so you'll find the right part of the code in a couple of minutes and rewrite it for your own needs.
In the book you suggested Rattle is mentioned 2 times! So it didn't help me, I couldn't find the actual information in Russian about this package, so I asked here.
You don't need the package, but elementary R language constructs to work with the package and data.
Speaking of English, there is a translator on Google. Copy and paste, that's it.
In the book you suggested Rattle is mentioned 2 times! So it didn't help me, I couldn't find the actual information in Russian about this package, so I asked here.
Here is an article on how to use Rattle https://www.mql5.com/ru/articles/1165
You will not believe, but this very article was the reason to download and install R. And, there is no answer to my question, which arose in connection with the method of SanSanych, namely downloading new data to check them on the trained models.
You will not believe, but this very article was the reason to download and install R. And, there is no answer to my question, which arose in connection with the method of SanSanych, namely the loading of new data to check them on the trained models.
Here are more articles with code examples
https://www.mql5.com/ru/users/vlad1949/publications
I read some of them three times and tried to figure out how those examples work. And then you can already make something of your own.
Oops... I have a lot of experience with this kind of traders, but I have a lot of experience in this kind of traders, and I have a lot of experience with this kind of traders, and I have a lot of experience with this kind of traders.) I've heard there are algotraders who have negative errors, it's in tough hedge funds, like Rentech, we traders are far away from them, but there's something to strive for, they use GPU and FPGA.
I remember, I remember. Alesha confidently stated, that 0.7 reliability is nonsense.
Alyosha:
They use GPUs and FPGAs.
These are high-frequency devices, the MO is nominally present there. And there patterns are obvious, especially through darkpools, you have time only to shoot orders. But expensive because of the infrastructure
There is no problem with gpu - Pytorch, even all the numpy matrix calculations can be transferred to the card
This is high-frequency, the MO is nominally present there. And there the patterns are obvious, especially through the darkpools, you only have time to shoot with warrants. But it's expensive because of the infrastructure.
Not so expensive because of the infrastructure. Slow down and everything will be OK. I have recently shown the system test - the deal is once every half an hour or so.
It's not that expensive. Lower speed and everything will be OK. I recently showed a test of the system - a transaction about once every half hour.
It's not even Low latency, hft is milliseconds or even microseconds, every extra meter of cable counts.
you have the usual scalping