Machine learning in trading: theory, models, practice and algo-trading - page 2282
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
Do R Connect
already did, but it didn't work outhttps://www.mql5.com/ru/forum/306688/page10#comment_11308158
Can you share the information?
2) display the error when the script stopped abruptly, like in python, with a link to the line and a description of the error and that would be by default in the log
3) add an integration with api telegram, at least a minimal set
4) allow the placing on the market advisers with webrequest, ie with the ability to request information from the outside
1) make a normal integration with sqlite3
2) display the error when the script stopped abruptly, like in python, with a link to the line and an error description, and by default send it to the journal
3) add integration with api telegram, at least a minimal set
4) allow to put on the market advisors with webrequest, ie with the ability to request information from the outside.
1) it has long existed as a standard DatabaseXXX functions, even a sqlite database browser is built into the editor, a detailed article about SQL in MQL5
2) the entire output of python scripts is written in logs in detail
3) this has long existed as separate solutions and does not apply to python library. I asked questions about the Python link. Instead of telegram it's convenient to use the regular pushers in cell phones
4) has always been available
it already happened, but it didn't work outhttps://www.mql5.com/ru/forum/306688/page10#comment_11308158
sadly....
the most popular product in the world after matlab for statistical calculations, call it a collective farm )))
The level is clear and so is the collective farm...
sadly....
the most popular product in the world after matlab for statistical calculations, call it a collective farm )))
The level is clear, and so is where the collective farm is...
The level is shown here: Statistical Distributions in MQL5 - Taking the Best of R and Making it Faster
Comparison of Calculation Speed
We have prepared scripts for measuring calculation time of density functions (pdf), probability functions (cdf), functions for computing quantiles and functions for generating pseudorandom numbers to compare the calculation speed of statistical functions.
The calculations were performed on an array of 51 values; for continuous distributions the function values were calculated within the range of 0 to 1, for discrete ones - from 0 to 50. The calculation time of the statistical functions in R has been measured using the microbenchmark library. The function calculation time in MQL5 was calculated using the GetMicrosecondCount() function. The TestStatBenchmark.mq5 calculation script can be found in the terminal_data_folder\MQL5\Scripts\UnitTests\Stat. The script for R and the results of calculation speed measurements are given in the Appendix.
The calculations were performed on Intel Core i7-4790, 3.6 Ghz CPU, 16 GB RAM, Windows 10 x64.
Calculation time measurement results (in microseconds):
calculation PDF (µs)
PDF (microseconds)
R/MQL5
CDF calculation (µs)
CDF calculation (µs)
R/MQL5
quantiles (µs)
quantile (µs)
R/MQL5
random numbers (μs)
random numbers (μs)
R/MQL5
Table 4. Time required to calculate statistical functions in R and MQL5 (in microseconds).
For R we took minimum time values, and for MQL5 we took average values (pdf_mean, cdf_mean, quantile_mean, random_mean).
The table shows that even under such conditions the calculation of functions of MQL5 statistical library is several times faster than R. On average, MQL5 calculates 3 to 7 times faster than R even taking into account that the compared versions of R functions are in fact written in C++.
In practice, the MQL5 language compiler has turned out to be much faster than C++ implementations of functions in R, which shows the high quality of our developments. When transferring programs from R to MQL5, you can get a significant speed-up and you do not need to use third-party DLLs.
MQL5 is now hundreds of times faster than Python, but with the introduction of native matrix operations as the basis for the language, we will see the speed of mathematics, that is several times faster than third-party Python modules.
1) has long been available as a regular DatabaseXXX functions, even the sqlite database browser is built into the editor, a detailed article about SQL in MQL5
2) the entire output of python scripts is written in logs in detail
3) this has long existed as separate solutions and does not apply to python library. I asked questions about the Python link. Instead of telegram it's convenient to use the regular pushers in cell phones
4) has always been available
The level is shown here: Statistical Distributions in MQL5 - Take the Best of R and Make it Faster
Speed is good, but far from decisive
When you titled your article "Statistical Distributions in MQL5 - Using the Best of R and Making it Faster" I
I swear I laughed all day long! You took three functions from millions of ones in R, invented them and rewrote them into MQL5 and that's it! ))) All of you have surpassed R!!!! Ahahaha )))) can't excuse me...
What about the fact that R has the largest database of any statistical methods, for any direction, thousands of libraries... When will you rewrite it?
What about the full list of any MO algorithms, hundreds of libraries... When are you going to rewrite it?
What about the CRAN database, where every day the world's latest products appear, which everyone wants and can try... When are you going to rewrite it? And most importantly WHY?
Do you know that I can replace a hundred µl in R with one line of code?
All you have in your arguments is speed, the speed of the three functions you took from R.
Did you know you can write any C++ function in R?
Eh, okay... Why am I making a big deal about it... It's useless anyway, you've already solved everything... and so have I...
One thing I ask - don't write anymore that "you take something better from R", you don't know what you're talking about...
Even if you rewrite a thousand functions on µl it will not be even a hundred percent of what is in R, what's the point of doing this foolishness?
Why not let people themselves port into this world of statistics, machine learning, signal processing, econometrics and much more...
Speed is good, but far from decisive.
When you titled your article "Statistical Distributions in MQL5 - Taking the Best of R and Making It Faster", I laughed the whole day!
I swear I laughed all day long! You took three functions from the millions in R, invented them and rewrote them into MQL5 and that's it! ))) All of you have surpassed R!!!! Ahahaha )))) can't excuse me...
We took and rewrote in pure MQL5 about 500 built-in R functions. That's what is the basic mathematical basis of R system. All the rest are third-party packages.
We have rewritten these functions faster and distribute them in the /include/math/stat source code. This shows the quality of our language and our developers.
We have made our choice as to who and how to support them:
When you're throwing questions and rants, I recommend that you look at your opponent, who:
If you're here to consume (no one here has seen your results in public), then behave decently and watch your language.
I don't consume anything here, I'm just communicating on the forum, if you ban me it will only make me feel better... So...
And these 500 functions that you rewrote, they're dead in the sense that they're like a ruler with which you can measure something, and this "something" is exactly the third-party packages to which you locked access
Would it make sense to make ZeroMQ native support?