Time to convert libraries to MQL5 - page 9
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 have some Matlab projects, will you clean them up?
You mean rewrite to MQL5? - Yes, why not.
I'm on it.
There is a working version of HHT, written in MQL4 using ALGLIB procedures (C++). I'm ready to share the algorithm and do the conversion to pure 5 as soon as ALGLIB port is available.
The HHT is only one of the methods, it can be easily integrated with some others.
Just need some feedback on who to send it to.
Well, ALGLIB is clear, MQ is being ported,
but is FANN being rewritten by someone? it seems to be the second most desirable bible.
Doesn't ALGLIB have what R has?
From the description, AGLIB cannot be compared with R - they are different level packages in favour of R. The arguments in favour are as follows:
1. R is freely distributable code without any restrictions
2. R is a lady in age (20 years), and if you take into account its commercial predecessor S, it's just old.
3. Russian localized version is available
4. At the moment R contains about 3500(!) packages, and it is originally oriented to statistics (that's the name of the package) and not mathematics (unlike ALGLIB).
5. There are five groups of dainty packages: statistics, econometrics, time series, finance (portfolios are included here), robust systems. Besides that there are filters, wavelets and splines, and many more for TS - just can't estimate. Most of these concepts in ANGLIB I did not see.
PS: neural networks are also available, so everything and for free.
5. All packages are accompanied by documentation
6. There is a huge (I don't know another such system) educational, methodological and scientific literature on the use of R packages in statistics, econometrics and time series. Т
7. R is becoming the language of description of algorithms in scientific publications on statistics econometrics and time series
8. Very well docked with C and C++. Although the R language itself is similar to LISP and may well compete with C++. There is a wide scope for coding enthusiasts, including writing very efficient code, including parallel computing.
9. A very elegant solution to code openness: any of the 3500 packages is always open source (R interpreter), nothing closed in the dll. This is the principle of the system. R itself has to be installed (minus) but its installation is primitive.
10. Has over 2 million users looking for bugs.
11. Those who want to try it can take the library from kodobase. I suggest you appreciate the elegance of accessing R. I hope to post a forecast indicator which is calculated in R.
Well, ALGLIB is clear, MQ is being ported,
but is FANN being rewritten by someone? it seems to be the second most desirable bible.
12. No porting (oh shit! Chief, it's all gone!)
I got the numbering wrong: two numbers 5. You don't have number 12 - you have 13.
Well, then it's more for you. I meant that because of the format R is practically not portable in principle.
And the wrapper used, firstly, allows you to use the product now, and secondly, it is very far from the priority selected for the full migration of the code to MQL5.
Unlike ALGLIB and FANN, which didn't please you.
Well then it's more for you. WhatI meant was that because of the R format it's almost impossible to port in principle. .....secondly it's very far from the priority chosen to fully port the code to MQL5
The code is open, what is the problem? Something you can, something you don't need, but personally I don't need anything at all. If we don't port anything, the problem with new R releases is solved.
And the wrapper we use, first of all, allows us to use the product at any time,
No doubt. But it is interesting to fudge indicators and scripts which are not in kodobase. I remember how much the MNCs chewed up, and there wouldn't be a problem here. Recently an article on nuclear assessment came out - again there wouldn't be a problem, and much wider than that.
Unlike ALGLIB and FANN, which didn't please you in any way.
Well, they did.
But if you fight for purity of idea, they did not please you. These are alien packages for trading. No, they have mate methods that are used and can be applied in trading. Take NS as an example. There are plenty of NS packages and they are applied. In econometrics packages it is in the classification section, its place is immediately clear and you can look through other classification methods without searching for it in other packages.
R is a system of interconnected and selected means for solving problems of econometrics and statistics applied in trading. Nothing is superfluous. A newcomer does not need to select, select and dock the tools. For a skilled person it may be easy, for example to type in Matlab, but for a newcomer it is a daunting task.
Unlike ALGLIB and FANN, which do not please you.
No, you're not. Compare the composition of R and the composition of these packages.