Machine learning in trading: theory, models, practice and algo-trading - page 674
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 need to more thoroughly investigate the testing question. In the meantime I dabble in genetic programming. They say that I can get decent results from it. Genetic programming gives me something similar to formulas. I think it should be easy to program them in MQL. https://smart-lab.ru/blog/269992.php The program is good for this purpose, but somewhat limited https://dev.heuristiclab.com/trac.fcgi/
It would be better to find a program for converting NS weights and structure into code...
Basically the same as NS. Same +-/* Exp. But NS has more useful functions for learning.
It would be better to find a program to convert NS weights and structure into code...
You're really something. MT - DLL - Python. and all the NS of the world is yours. And there is no need to write-convert anything.
The second option is even easier: MT - files - Ram -Disk - Python (R). Performance is as fast as the DLL. And in general you don't have to write anything. And again, the NS of the whole world is yours.
I do not understand what is going on. Why would anyone want to convert NS into code, MQL, etc.?
They know how to make life difficult for themselves.
You're really something. MT - DLL - Python. and all the NS of the world is yours. And there is no need to write-convert anything.
The second option is even easier: MT - files - Ram -Disk - Python (R). Performance is as fast as the DLL. And in general you don't have to write anything. And again, the NS of the whole world is yours.
I do not understand what is going on. Why would anyone want to convert NS into code, MQL, etc.?
They know how to make life difficult for themselves.
Train on R. Convert it to MT code - throw it on UPU (which has nothing but MT) and forget it.
In a week, train again and again throw the retrained network in the form of a code in MT.
However, this is again tuning - first you need to get a working engine, and then you can do it.Gentlemen! I propose to spit on the current configuration of this branch and powerfully revive interest in it.
A certain person, abandoning the momentary pursuit of cash, should take the courage to make the NS work in forecasting mode, and in no other way.
By prediction we mean the classical understanding - predicting the object's behavior in the future, say, in 1 day or 1 hour. With graphs and comments to them.
This will be interesting. And to this person, honor and respect.
Gentlemen! I propose to spit on the current configuration of this branch and powerfully revive interest in it.
A certain person, abandoning the momentary pursuit of cash, should take the courage to make the NS work in forecasting mode, and in no other way.
By prediction we mean the classical understanding - predicting the object's behavior in the future, say, in 1 day or 1 hour. With graphs and comments to them.
This will be interesting. And this person will be honored and respected.
Before solving a problem, you should at least find out whether it has a solution).
I sincerely believe that it does not. And, it seems to me, that it is almost proved by previous attempts of market forecasting.
Before solving a problem, you should at least find out if it has a solution.)
I sincerely believe that it does not. And it seems to me that this has already been almost proven by previous attempts at market forecasting.
Then we need a convincing proof of this fact with several variants of input data: prices, increments, sums of increments, etc.
Then you need a convincing proof of this fact with several variants of the input data: prices, increments, sums of increments, etc.
I don't need any.) The fact that it didn't work out for anyone reliably is already circumstantial, but proof.
In general, you cards in your hands, and forward for the orders).
Then you need a convincing proof of this fact with several variants of input data: prices, increments, sums of increments, etc.
The convincing proof is that nobody in this thread has ever managed to earn with the help of NS, with or without forecasting
Proceed from this fact ))
I, for example, have the last variant of idea, for which I'll spend the remaining 1.5 weeks and will forget about this theme )) I'll show the results
A convincing proof is that no one in this thread has ever been able to earn with the help of NS, with or without forecasting.
Proceed from this fact ))
I have, for example, the last variant of the idea, for which I'll spend the remaining 1.5 weeks and will forget about this subject )) I'll show the results