Machine learning in trading: theory, models, practice and algo-trading - page 663
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
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. If you want to use MT terminal, let it solve the tasks of the terminal - market data and orders-transactions.
For example, in order to display all predictions of the NS on the training area and forward on a chart and visually assess whether I want the NS or it has learned some bullshit, which accidentally turned out to be profitable on a certain part of history.
Why not in R? It has extra class graphics.
For example, to display all that has predicted the NS on the training area and forward to a chart and visually assess - what do I want from the NS, or it has learned some bullshit, that accidentally turned out to be winning at a particular section of history.
I'm not trying to change your mind. The same in R and other modeling software is done much easier. In addition, everything can be quickly changed, calculated, output additional graphs or pieces of graphs, calculate any statistics, etc. And all this is literally in 5 minutes.
In MT, of course, you can do it too, but it's much harder and longer.
Why not in R? It has extra class graphics.
I'm not trying to change your mind. The same in R and other modeling software is done much easier. Besides everything can be quickly changed, calculated, output additional graphs or chunks of graphs, calculate some statistics, etc. And all this is literally in 5 minutes.
In MT, of course, you can do it too, but it's much harder and longer.
It is more convenient to see everything in the terminal. You can put an arrow right on the desired bar. Scaling and scrolling - everything will be there. No drawing will show so many arrows legibly. Although, 100000 arrows will slow down the terminal too... Well, you can draw them only on the visible part of the window and redraw them when scrolling.
Everything is more familiar to see in the terminal. It is possible to place the arrow right on the desired bar. Scaling and scrolling - everything will be there.
At first I also did so, but then I gave up - everything is much more convenient in R, it has a chart as such and if you start to analyze the result, you can draw a lot of things. Even the simplest things in the terminal cause great difficulties. For example, to draw the difference between the model fit and the quote.
By the way, I completely abandoned the use of DLL, at least in software development. All information goes back there through text (CSV) files and RAM-Disk. The speed is acceptable even for scalping-piping > 1.5Gb/s. Advantage is quick (minutes) change-add any info to transfer channel.
And DLL can be done later, when all exchange protocols will be settled, and already for ready software.
At first I also did so, but then I gave up - everything is much more convenient in R, it has a chart as such and if you start to analyze the result, you can draw a lot of things. Even the simplest things in the terminal cause great difficulties. For example, to draw the difference between the model fit and the quote
At first I did it too, but then I gave it up - everything is much more convenient in R, it has a chart as such and if you start analyzing the result, you can draw a lot of things. Even the simplest things in the terminal cause great difficulties. For example, to draw the difference between the model fit and the quote
+1
By the way, I completely abandoned the use of DLL, at least in software development. All info goes back there via text (CSV) files and RAM-Disk/ The speed is acceptable even for scalping-piping > 1.5Gb/s. Advantage - fast (9 minutes) change-add any info to transfer channel.
And DLL can be done later, when all the exchange protocols will be established, and already for ready-to-use software.
By the way, I completely abandoned the use of DLL, at least in software development. All information goes back there through text (CSV) files and RAM-Disk. The speed is acceptable even for scalping-piping > 1.5Gb/s. Advantage is quick (minutes) change-add any info to transfer channel.
And DLL can be done later, when all exchange protocols will be settled, and already for ready-to-use software.
You are not the first time you write about the files.
And how is the problem of readiness polling solved?