Machine learning in trading: theory, models, practice and algo-trading - page 150

[mytarmailS]

Andrey Dik:

I showed you, several times already. Ask Azulenko and Perevenko, they know what we are talking about. And read your posts again, it will be clear, maybe...

Well all these attacks began after my post, so the claims are for me ... So where did I say something bad about mql?

A quote please...

[Andrey Dik]

mytarmailS:

Well all these attacks started after my post, so there are claims to me... So where did I say I was criticizing mql?

quote please...

The first was your post:

This is the forum on trading, automated trading systems and testing trading strategies

Machine learning: theory and practice (trading and beyond)

mytarmailS, 2016.10.09 13:46

I don't trade forex, I'm not even familiar with metatrader :)

I would like to experiment with the profile in its different forms in the R-ka, it is just not clear how to build a distribution of two vectors

I don't need volumes, it's just a start, you can put almost anything into the profile, but it has to be done with price, and that means two vectors

After that, it will be like this:

Forum on trading, automated trading systems and trading strategies testing

Machine learning: theory and practice (trading and beyond)

MytarmailS, 2016.10.09 14:44

I don't know how to do it :) I'm not familiar with MQL

I'm not familiar with MQL and I've got such problems, that even I (who's not a programmer) understand, that there are no built-in functions here, you have to write everything yourself... The fact that you translated a dozen statistical functions from R is certainly good and even commendable, but understand this is not even a drop in the ocean compared to the capabilitiesof R, it is constantly evolving, every day there are new libraries, not functions but libraries, how can you keep up with this?

So, not knowing MQL at all, you declare that MQL doesn't compare with R, forgetting explicitly that MQL is a full-fledged independent language, that already has a lot of various stat and matrix libraries, and at this very moment the standard libraries are being created. While R by and large is just a collection of libraries and functions, it can hardly even be called a language. In addition, there are innumerable different ready-made open source libraries in C and C++, that are ported to MQL with minimal effort. Visit kodobase or kodeproject (I often go there when I'm short of time and need a solution very quickly) or similar and use already available tools, which you can modify if necessary.

[mytarmailS]

Andrey Dik:

First there was your post like this:

After that it was like this:

So, not knowing MQL at all, you declare that MQL is incomparable to R, forgetting explicitly that MQL is a full-fledged independent language, for which many different stat and math libraries have ALREADY been written, and at the present moment the libraries are being written as a standard part of delivery. While R by and large is just a collection of libraries and functions, it can hardly even be called a language. In addition, there are innumerable different ready-made open source libraries in C and C++, that are ported to MQL with minimal effort. Visit kodobase or kodeproject (I often go there when I'm short of time and need a solution very quickly) or similar and use already available tools, which you can modify if necessary.

I agree with what I wrote, but still, where in these quotes did I say anything bad about mql?

[Andrey Dik]

mytarmailS:
I agree with what I wrote, but where in these quotes did I say anything bad about mql?

Are you being unconscious? I've already written the posts. Shall we continue to play the game "Where? - There! Where? - There!"? It would be much more productive for you, and then happily share with everyone here what you've found there.

Saying "The fact that you translated a dozen of statistical functions fromR is certainly good and even commendable, but understand that this is not even a drop in the ocean compared to the capabilitiesof R,it is constantly evolving every day new libraries appear, not functions but libraries, how can you keep up with this? and why? "You can ' t even imagine how many things are already done and available in C/C++/MQL source code.

[Dr. Trader]

mytarmailS:
I agree with what I wrote, but still, where in these quotes did I say anything bad about mql?

This is not how logic works in Andrei, you have to talk to him like a kindergartener, only straightforward sentences of a couple of words, otherwise he makes up the unknown and confuses himself.

The fact that you don't have a single bad word about MQL - that doesn't mean anything. You have written that R has a lot more possibilities. He doesn't mean "the number of modeling and data analysis libraries in R is greater than the number of modeling and data analysis libraries in MQL",
, but "MQL doesn't have any libraries, everything sucks". Teenage maximalism.

In general it is useless to explain him anything, no logic, no common sense, no adequate answers. Maybe he is a neural network trained on posts from this forum, I've seen such projects with recurrence network for writing syntactically correct but absolutely meaningless texts.

[Renat Fatkhullin]

Please stop with the accusations.

Every language has its place. R is great for interactive research. This is my second day exploring it (I read the book before) and it really looks like a powerful debugger with visualization of the guts.

Working with R has immediately revealed our weaknesses:

• MQL5 has few powerful functions for frequent operations. For many things you have to write microcode. In the next two builds we will roll out dozens of new functions for complex operations in a single call.
• We need more math functions. We've already released the first version of the R function analogue in beta and will now take it further by adding vector variants.
• We need a simple and powerful graphical library with functionality like graph packages in R. We will create it with an eye on R.

What are we doing it for?

We have released the first algorithmic trading platform with the MQL language back in 2001. Each time we increased its possibilities, but the mathematical algorithm left much to be desired. We were developing the analysis, data access, tester, distributed calculations, and then we started to sell our products.

And then it became clear that most of the solutions were stuck in a vicious circle of theanalysis, indicators and fitting. We need to let the developers get to the next level of mathematical capability.

That is why we started extending mathematical libraries in MQL5 some time ago and also released in beta Alglib, Fuzzy and Stat. They will make it easier to transfer worked-out models from other systems to MQL5 and raise the class of created analytical solutions for the Metatrader 5 platform.

In the next 2 months you will see the progress we will make in developing the mathematical environment.

We welcome and welcome discussion of complex mathematical packages as well as articles on them. Write and send requests for articles to Rashid Umarov (Rosh). Our objective is to encourage and educate traders in more sophisticated methods, not to fence them in our own MQL5 world.

Of course, we are and will be defending our language and platform from attacks, but we are also working to develop them. So everything will be fine.

[mytarmailS]

Andrey Dik:

Are you being unconscious? I've already written the posts. Shall we continue to play the game "Where? - There! Where? - There!"? It would be much more productive for you, and then happily share with everyone here what you've found there.

Saying "The fact that you translated a dozen of statistical functions fromR is certainly good and even commendable, but understand that this is not even a drop in the ocean compared to the capabilitiesof R,it is constantly evolving every day new libraries appear, not functions but libraries, how can you keep up with this? and why? "You have no idea how much has already been done and is available in source code for C/C++/MQL.

I'm explaining it to you my inconsiderate interlocutor :) since everything is very clear to everyone

In the context of communication with the administrator of the phrase

"The fact that you translated a dozen of stat. functions fromR is certainly good and even commendable, but understand, this is not even a drop in the ocean compared to the capabilitiesof R,it is constantly evolving, new libraries appear every day, not functions but libraries, how can you keep up with this? and why? it just needs to be used... "

meant that - if you think that taking a dozen of basic functions from R, you'll take the best ofR (and these were the words in the title of the article, to which the administrator sent me) , it will bea drop in the ocean compared to the capabilitiesof R

What does this have to do with your mql? What are you babbling about, Andrei? Come to your senses at last! or be a man and admit that you just didn't read carefully and saw what you wanted to see and not what is real ...

Dr.Trader:

Maybe he's a neural network trained on posts from this forum, I've seen such projects with a recurrence network for writing syntactically correct but completely meaningless texts.

Shit, maybe really? :)

Renat Fatkhullin:

Every language has its place.

I absolutely agree...

If I research the market and I want, say, to check some idea, "an example is taken from the ceiling" 1) cluster prices using special method for BP (DTW), then 2) train a hidden Markov model, 3) check the results by crossvalidation method

If I try to use R-type it will take 15-30 lines of code and 15 minutes of my time, with mql it will probably take more than 2000 lines of code and a week to debug, and since market ideas work in 5% of cases, R-type has some advantages over mql in the research area, but when the time comes to sell, then of course R-type is not used, and mql or its analogues that are designed for the very development of trading robots should already be used

So you're right - Every language has its place.

R is for market research

mql - for writing trading robots

==============================

But I don't know why you need to copy what's already written in R, so you can copy 20% of R's basic functions, distributions, correlations, approximations, interpolations, etc.

but that would only be 20% of the R-ki base, and the R-ki base itself is 3% of the R-ki as a whole...

A week ago I experimented with Markov models (that's machine learning), a few weeks ago I experimented with (SSA and wavelets) it's spectral analysis, etc.. You won't port to every user's idea some library from R-ki?

Why not just make a good link between R-ka and mql, and let people decide by themselves what they need, it's 1000 times less work for you and every user will find what they are looking for and not only what you ported.

[СанСаныч Фоменко]

mytarmailS:

R - for market research

Justify.

I don't see any problem using R on real accounts. I use rf at the moment. Expanding the use of R is my individual limitations, but in no way R.

PS.

Made an attempt to replace rf with trees in alglib - broke after half an hour: impossible to understand ash at all. And that's assuming I have a very decent understanding of rf thanks to R.

[mytarmailS]

SanSanych Fomenko:

Justify.

I mean trading, position management, stops, takeaways .....

If you have an mql you do it all in one line, with R-ka you have to write everything yourself and you don't need mql or analogues

[СанСаныч Фоменко]

mytarmailS:

I mean trading, position management, stops, takeaways .....

With mql it's all done in one line, with r-ka you have to do it all yourself and it's not needed here only mql or analogues

I agree.

MT is a trading terminal and to replace it with R is pure madness.

In addition, there is a language that allows you to fundamentally reduce the risks of trading. I have a fairly large amount of code.

I don't see any alternatives to replace MKL languages

But the trading decision block, especially considering that I don't recognize TA at all, R is in all its glory.

Of course R is beyond competition at the stage of development:

• interpreter,
• very compact code with a lot of meaning,
• hundreds of packages for pre-processing of cotier, the simulation itself and the evaluation of these models,
• low level of bugs,
• extensive literature...

In general, paradise in the development of decision-making blocks for positions in the market.