Machine learning in trading: theory, models, practice and algo-trading - page 148
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
Men and women, stop discussing personalities and move on to discussing machine learning. Personalities, if one so desires, can and probably should be discussed only in private.
Quite right, Yura. And it would be desirable to lead this very discussion in such a way, that the experiments could be repeated by those, for whom this wonderful resource was created - the MT users and MQL programmers.
I'm honestly afraid to think how much code I'll have to write to repeat in MQL, not even the preparation of training examples and training using some method, but:
I think it's called a kick in the teeth of MQL, isn't it?
who don't know a thing about MQL and are scandalized by the language they don't know
It is enough just to read documentation to understand what tool is better suited to the task. You need to look at the developed application from an architectural point of view - find descriptions of functions, search for appropriate frameworks or libraries and draw a block diagram of the program logic for one language and then for another. Compare the schemes, and the easier language wins. To do this it is enough to understand the general principles of building programs in these languages, which interventions are there, what types of data etc. MQL is simple, you can consider it as a framework for c++ or java with functions for trading and working with time series and indicators. If you know at least javascript you can do something in mql, all you need is to know trading functions.
Example. I don't know PHP, but I have an idea to make a trading Expert Advisor with it. I can find a general-purpose scripting language that is widely used for web applications development. I have searched for frameworks to work with time series, for connecting to stock exchanges and for modeling - have found nothing. So it can be concluded that the trading Expert Advisor in php is not worth doing, we need a lot of self-made libraries.
Also we can evaluate the use of mql for creating EAs with very complex models and logic - if you google the keywords did not give anything, then it is better not to try and use a language where this is done easily and effortlessly.
Regarding R - there are no libraries specifically for forex trading. That's why I'm exporting ohlc and indicators and all trading operations in my EA in mql, and training of trading models and decision making in R. I suggested to use R as an example and I've already seen it with my mql Expert Advisor, I've always wanted to trade with R. The only crutch is synchronization of these two programs through files, everything else is standard and works fine.
They are a parasite on the MQL community
You're talking about parasitism again. The parasites usually take something away from the victim, I don't see how they can take something away from MQL. If you want to show the criteria of parasitism in MQL, please provide them.
There is a very piquant situation with "parasites.
It may seem from the posts of our tireless bicycle inventor thatVladimir Perervenko is classified as a "parasite" on the grounds that he publishes articles that have no code on the mcl.
The funny thing is thatVladimir Perervenko as well as other authors of articles, including me, sold the exclusive right to publish their articles to metaquotes for money. That is,Vladimir Perervenko HAS NO RIGHTS TO PUBLICATE his articles. It is on the terms of the transfer of publishing rights to the methaquots that articles appear on this site. And all the articles that we see on this site are the methaquots exercising their purchased right to publish someone else's articles.
So who exactly has Andrew Dick labeled as a "parasite"?
The use of third-party (open source!) software to assist trading in the MT trading platform is not just bad or neutral, it is even good and integrates MT into the statistician environment.
MT's growth as a platform for sophisticated modeling is a plus.
The fact that MT allows articles to be written using third-party software code is a definite plus for the company.
Men and women, stop discussing personalities and move on to discussing machine learning. Personalities, if you're so eager, can and probably should be discussed only in private.
T allows you to write articles using third-party software code, which is a definite plus for the company.
Not only does it allow it, but it also buys the publishing rights and publishes it.
For example, my articles are not only published, but also translated into other languages and published.
I should note that metaquotes policy regarding articles in the end makes this site sharply different from any other forex sites.
It is enough just to read the documentation to understand which tool is best suited to the task. You need to look at the application being created from an architectural point of view - find function descriptions, search for appropriate frameworks or libraries, and draw a block diagram of the program logic for one language, then for another. Compare the schemes, and the easier the language, the better. To do this it is enough to understand the general principles of building programs in these languages, which interventions are there, what types of data etc. MQL is simple, you can consider it as a framework for c++ or java with functions for trading and working with time series and indicators. If you know at least javascript you can do something in mql, all you need is to know trading functions.
Example. I don't know PHP, but I have an idea to make a trading Expert Advisor with it. I can find a general-purpose scripting language that is widely used for web applications development. I have searched for frameworks to work with time series, for connecting to stock exchanges and for modeling - have found nothing. So it can be concluded that the trading Expert Advisor in php is not worth doing, we need a lot of self-made libraries.
Also we can evaluate the use of mql for creating EAs with very complex models and logic - if you google the keywords did not give anything, then it is better not to try and use a language where this is done easily and effortlessly.
Regarding R - there are no libraries specifically for forex trading. That's why I'm exporting ohlc and indicators and all trading operations in my EA in mql, and training of trading models and decision making in R. The only crutch is synchronization of these two programs via files, everything else is standard and works great.
There you go again about parasitism. Parasites usually take something from the victim, I don't see how you can take something from MQL. The criteria of parasitism in MQL, please provide them.
Do you understand what you just said? "The MQL language is not suitable for creating Expert Advisors with very complex models and logic..." That's what Azulenko, Perevenko, and Fomenko are saying. These are people who don't know MQL and don't know its features, who haven't even tried to write something serious enough in MQL...
They don't say it's impossible to work in MQL, they just create.
There's no need to criticize, you know? Blaspheme, blame and calumniate.
I gave the example of refugees for a reason. Live, develop, do your business, but respect the place where you live and the community in which you are, while benefiting this resource by developing and multiplying the wealth in the form of codes and scientific and technological solutions and achievements. Otherwise you are just parasites.
Otherwise, you're just parasites.
I wonder, will they start to ban the rude ones, not all of them, but the ugly ones?
This is where you start reading:
You don't know the MQL language, you don't know the terminal, how can you judge anything at all?
PS. It's a pity, but there are so many users on this forum who are parasitizing on MT ecosystem. But it also shows the huge popularity and sophistication of the platform, which attracts the smartest people to research in MQL, and, of course, attracts parasites, how could it be without them...
Where I reasonably noted that in order to assert something, you must first KNOW the subject of the statement.
This was followed by the statement of Perevenko, who used to speak about "state-of-the-art methods in R", but who technically merged when I suggested comparing his algorithms in R with mine in MQL:
The "Pioneers" were "show me, prove it, let me try it".
The "Komsomol members" were - "forbid to publish..."
Now there are "environmentalists" as well - the purity of the "ecosystem" is sacred...
Indeed - stupidity is boundless.
Well, which one of us is stoned? You are the stubborn R fanatics who can't see further than their nose, busting their foreheads for R, but who are unable to prove the superiority of those very "newest cutting-edge methods in R" over MQL.