Attachments: MQL4 -> R-Project - Interface Library

MQL4 -> R-Project - Interface Library

#1
First Post: Edited Nov 17, 2010 3:08am Oct 11, 2010 1:20pm | Edited Nov 17, 2010 3:08am

7bit
Joined Mar 2009 | Status: Trader | 1,231 Posts

[Keywords: mt4R.dll, mt4R.mqh, R, Metatrader]

Hi,

maybe some of you want to experiment with my MT4 -> R interface that allows one to start as many R sessions (usually one per running EA), transfer data forth and back and call R functions. The R sessions will be run as separate processes in the background but the API towards mql4 are synchronous function calls, no ugly text files writing and polling. Communication goes through stdio pipes to and from each of the R processes.

In this .zip file are the needed files (mt4R.mqh and mt4R.dll) for running this thing. The .mhq file also contains all API documentation and other needed information.

Attached File(s)

mt4R-1.3.0.25.zip 101 KB | 5,568 downloads | Uploaded Nov 6, 2010 5:40pm

And optionally for those interested in how the DLL works below is the dll source code (it needs FPC 2.4.2 or newer to compile. You should install a current snapshot of Lazarus/FPC with a 2.4.3 compiler, you will find a current snapshot installer (IDE and Compiler in one package) for win32 on ftp://ftp.freepascal.org/pub/lazarus/snapshots/).

Here is the source:

Attached File(s)

mt4R_source-1.3.0.25.zip 19 KB | 3,715 downloads | Uploaded Nov 6, 2010 5:40pm

The above files are updated to the latest version. Forget the preview files in post #2, they are old.

#2
Oct 13, 2010 3:37pm Oct 13, 2010 3:37pm

7bit
Joined Mar 2009 | Status: Trader | 1,231 Posts

new version:

added RAssignMatrix()
added RAssignFactor()

and a few other things.

Attached File(s)

mt4R_preview_2010_10_13.zip 92 KB | 2,929 downloads

mt4R_preview_source_2010_10_13.zip 9 KB | 2,263 downloads

#3
Oct 17, 2010 11:03am Oct 17, 2010 11:03am

7bit
Joined Mar 2009 | Status: Trader | 1,231 Posts

Bump & new version.

I updated the files in posting #1, please ignore posting #2, it is outdated but I cannot delete or update it anymore

* renamed and changed some API functions for more consistency
* header file will now check for correct DLL version
* improved the mechanism how vectors are transferred to R
* got rid of the Sleep() in TRConsole.GetOutput(), I realized that I can do blocking IO with the stdio pipes and where it previously spent up to 20 milliseconds waiting for the response it takes only a few dozen microseconds now from request to response and you get the full throughput, no polling and sleeping, the CPU is always at 100% in the backtester.

Now I want to see some EAs or indicators from you making use of this.

#4
Oct 18, 2010 1:53pm Oct 18, 2010 1:53pm

7bit
Joined Mar 2009 | Status: Trader | 1,231 Posts

Attached Image (click to enlarge)

Click to Enlarge

Name: 2010-10-18-204653_1023x664_scrot.png
Size: 105 KB

PS:
Latest version might contain a bug: I have sporadic crashes since I removed cmem from the uses clause. Might have to add it back.

#5
Oct 18, 2010 3:10pm Oct 18, 2010 3:10pm

7bit
Joined Mar 2009 | Status: Trader | 1,231 Posts

updated the files in post #1

#6
Oct 20, 2010 6:07pm Oct 20, 2010 6:07pm

7bit
Joined Mar 2009 | Status: Trader | 1,231 Posts

updated the files in post #1

API changed, therefore we have a change in the version number 1.1 -> 1.2 and a new version of the .mqh is needed.
new: RExecuteAsync() to execute code in the background and not wait
new: RIsBusy() usually called at the beginning of start() to detect whether the previous call to RExecuteAsync() during an earlier tick is still running.

These two new functions are useful for making indicators. Unlike EAs and scripts an indicator runs in the GUI thread of MT4 and may not block. The start() function must immediately return. The suggested pattern now is to call RIsBusy() on the beginning of start() and do nothing if this returns true and only if R is idle you can use R functions and the last R call in the start function should be the one that starts the long running number-crunching task and should be done with RExecuteAsync() instead of RExecute(). After that the start() function can return while R will continue running in the background. The next ticks will then use RIsBusy() to check whether it has finished.

If one of the R functions is called while it is still executing another command it will simply wait until it is idle again and then execute the command. Only one R comand can execute at a time and there is one lock per R session to ensure this. There will be no errors (only debug messages), everything will still happen in the correct order, only your start() function might not return quickly if it is forced to wait for the lock, so be careful to intelligently use RIsBusy() and intelligently organize the control flow to go through start with minimal blocking.

If you do EAs then all this is not necessary (or even contraproductive), here the start() function may block as long as it wants and for the strategy tester this is even needed. RExecuteAsync() is only a workaround for indicators and MT4's ugly design flaw of running all indicators in the main thread, here you need this to do long running tasks asynchronously to make start() always return quickly.

Tomorrow I will post a little example indicator.

#7
Oct 21, 2010 2:13am Oct 21, 2010 2:13am

hammer083
| Joined Sep 2010 | Status: Trader | 45 Posts

Hello,

Can you please share your auto-regression indicator for MT4?

Thank you,

hammer083

#8
Edited 9:12am Oct 21, 2010 6:59am | Edited 9:12am

7bit
Joined Mar 2009 | Status: Trader | 1,231 Posts

Here it is: (save it as an indicator in the indicators folder)

Inserted Code

[color=Blue]#property copyright[/color] [color=Green]"© Bernd Kreuss"[/color]


[color=Blue]#property indicator_chart_window[/color]

[color=Blue]#property indicator_buffers[/color] [color=Green]1[/color]
[color=Blue]#property indicator_color1[/color] Red
[color=Blue]#property indicator_width1[/color] [color=Green]2[/color]


[color=Blue]#include[/color] [b]<mt4R.mqh>[/b]

[color=Blue]extern int[/color] order = [color=Green]200[/color];
[color=Blue]extern int[/color] back = [color=Green]500[/color];
[color=Blue]extern int[/color] ahead = [color=Green]20[/color];

[color=Blue]int[/color] R;
[color=Blue]double[/color] buf_prediction[];

[color=Blue]int[/color] init(){
   SetIndexBuffer([color=Green]0[/color], buf_prediction);
   SetIndexStyle([color=Green]0[/color], DRAW_LINE);
   SetIndexShift([color=Green]0[/color], ahead);
   R = RInit([color=Green]"C:/Programme/R/R-2.11.1/bin/Rterm.exe --no-save"[/color], [color=Green]2[/color]);
   Comment[color=SeaGreen]([color=Green]"history: "[/color][/color] + back +[color=Green] " bars, method: OLS, order: "[/color] + order);
}

[color=Blue]int[/color] deinit(){
   RDeinit(R);
}

[color=Blue]int[/color] start(){
   [color=Blue]double[/color] hist[];
   [color=Blue]double[/color] pred[];
   [color=Blue]int[/color] i;
   
   [color=Blue]if[/color] (RIsBusy(R)){
      [color=Gray]// last RExecuteAsync() is still not finished, do nothing.[/color]
      [color=Blue]return[/color]([color=Green]0[/color]);
   }
   
   [color=Blue]if[/color] (RGetInteger(R, [color=Green]"as.integer(exists('model'))"[/color]) == [color=Green]1[/color]){
      [color=Gray]// there exists a model (the variable is set). 
      // This means a previously started RExecuteAsync() has finished. 
      // we can now predict from this model and plot it.[/color]
      RAssignInteger(R, [color=Green]"ahead"[/color], ahead);
      RExecute(R, [color=Green]"pred <- predict(model, n.ahead=ahead)$pred"[/color]);
      ArrayResize(pred, ahead);
      RGetVector(R, [color=Green]"rev(pred)"[/color], pred, ahead);
      [color=Blue]for[/color] (i=[color=Green]0[/color]; i<ahead; i++){
         buf_prediction[i] = pred[i];
      }
   }
   
   [color=Gray]// make a (new) prediction
   // move some history over to R   [/color]
   ArrayResize(hist, back);
   [color=Blue]for[/color] (i=[color=Green]0[/color]; i<back; i++){
      hist[i] = Close[i];
   }
   RAssignVector(R, [color=Green]"hist"[/color], hist, ArraySize(hist));
   RExecute(R, [color=Green]"hist <- rev(hist)"[/color]);
   
[color=Gray]   // crunch the numbers in the background and return from the start() function
   // RIsBusy() in the next ticks will tell us when it is finished.[/color]
   RAssignInteger(R, [color=Green]"ord"[/color], order);
   RExecuteAsync(R, [color=Green]"model <- ar(hist, aic=FALSE, order=ord, method='ols')"[/color]);
   [color=Blue]return[/color]([color=Green]0[/color]);   
}

This will keep the CPU busy all the time but it will not block the MT4 GUI. The start() function will always immediately return, the only long running command is the call to ar() and this is done in RExecuteAsync() and the whole start() function is written in a way that this is the very last R call before we return (this is important), therefore all the ifs and the somehow twisted control flow (plot the old prediction before we fit a new model).

This indicator is an experimental quick & dirty prototype that I wrote for testing the async execution. Maybe some code could be added to make sure it executes only on bar open but then it should loop from 1 to back and not from 0 to back-1 because on bar open Close[0] and Close[1] are from the same time and not 1 hour apart and also the indicator shift must be reduced by 1 bar to predict the current close and not the next.

Here is how it looks when attached to an eurusd H1 chart (It seems it works best on higher time frames (H1 and longer) that seem to show strong periodicity, and eurusd seems to work very well):

Attached Image (click to enlarge)

Click to Enlarge

Name: 2010-10-20-052108_1019x521_scrot.png
Size: 42 KB

Attached Image (click to enlarge)

Click to Enlarge

Name: 2010-10-20-155318_1021x523_scrot.png
Size: 42 KB

Attached Image (click to enlarge)

Click to Enlarge

Name: 2010-10-21-133830_1023x523_scrot.png
Size: 44 KB

#9
Oct 21, 2010 9:35am Oct 21, 2010 9:35am

7bit
Joined Mar 2009 | Status: Trader | 1,231 Posts

I want to repeat some things some of them are mentioned in the .mqh file or are obvious (at least to me) but it won't hurt emphasizing them again:

First and foremost you need DebugView.exe to view the debug output. Without it you are blind! Don't even try to figure out what is going on without looking at the output of DebugView (while mt4r is set to debug level 2. Later you can set it to 0). It will not print anything into the MT4 log, the only exception is if it cannot load the DLL at all, this will be seen in MT4's log!
If you want to experiment with an interim result directly in the R console you can RExecute(R, "save.image('c:/some/path/some/file.R')"); to save the current environment and then fire up a new R console and load('c:/some/path/some/file.R') to restore all variables like they were when you called save.image() and then experiment with them in the interactive R console.
If you want to use plot() from within mql4 then make sure you also set the default device in init (like in the screenshot in the 4th posting). Note that the plot window will freeze if you do not execute at least one dummy command on every tick to keep R's event loop alive. It was never meant to be used this way. You can also plot to a file, see the R manual.
if you forget a closing ) or something like that it might freeze and wait for more input (which won't come), this might freeze MT4 if you are making an indicator, killing Rterm.exe in the task-manager will un-freeze it again. Start with EAs or scripts first until you are comfortable with it, they are allowed to freeze and block and you can still remove them from the chart.
Did I mention you should look at the output of DebugView.exe?

#10
Edited 10:10am Oct 21, 2010 9:37am | Edited 10:10am

Leche Mocha
| Joined Aug 2009 | Status: "Protect yo neck"- WuTang Financial | 854 Posts

7bit, Congrats on your cool indicator. It would be great if it could be put into a downloadable mq4 or EX4 format.

#11
Edited 10:37am Oct 21, 2010 10:07am | Edited 10:37am

stevegee58
Joined Oct 2005 | Status: Pip Slappa Extrordinaire | 1,013 Posts

Damn Bernd, this thing is cool. I'd been avoiding learning about R because I have 8 million other things I'm looking at. But this autoregression indicator was the perfect excuse to install and try it out.

I just had to change the path to the R installation to make it work.

BTW, I've adopted your common_functions.mqh for my own use and added a new function for stop-based position sizing if you're interested.

I've observed that on the H1 timeframe EURUSD has been very reliable using a linear regression channel going back to the beginning of September. I'm sure it's not a coincidence that you've been observing the same thing with autoregression.

You are in a maze of twisty little passages, all alike.

#12
Oct 21, 2010 6:35pm Oct 21, 2010 6:35pm

noobie1
| Joined Jan 2008 | Status: Trader | 28 Posts

Can't seem to get the indicator to work... tried to compile it and it can't open it... not sure what I'm doing wrong... any help?? Thanks

#13
Oct 21, 2010 6:52pm Oct 21, 2010 6:52pm

stevegee58
Joined Oct 2005 | Status: Pip Slappa Extrordinaire | 1,013 Posts

Put the dll in the experts/libraries directory, put the mqh file in the experts/include directory. In 7bit's pasted code, make sure that the following line points to where *you* installed R:

Inserted Code

 R = RInit("C:/Programme/R/R-2.11.1/bin/Rterm.exe --no-save", 2);

Also, don't forget to install R.

You are in a maze of twisty little passages, all alike.

#14
Oct 21, 2010 7:52pm Oct 21, 2010 7:52pm

noobie1
| Joined Jan 2008 | Status: Trader | 28 Posts

Thanks Steve... got it to open and compile with no errors... got it to open the dialog box on the chart, but it won't go any further than that. Heh... told you I was a noob!!

#15
Oct 21, 2010 9:10pm Oct 21, 2010 9:10pm

stevegee58
Joined Oct 2005 | Status: Pip Slappa Extrordinaire | 1,013 Posts

Are you on Win XP, Windows 7 or Vista?

You are in a maze of twisty little passages, all alike.

#16
Oct 21, 2010 9:43pm Oct 21, 2010 9:43pm

noobie1
| Joined Jan 2008 | Status: Trader | 28 Posts

#17
Oct 21, 2010 9:58pm Oct 21, 2010 9:58pm

stevegee58
Joined Oct 2005 | Status: Pip Slappa Extrordinaire | 1,013 Posts

Well then it's not what I was thinking LOL.

Make sure the slashes in the directory path for the R executable in the indicator source are pointing like "/" instead of "".

You are in a maze of twisty little passages, all alike.

#18
Oct 22, 2010 8:05am Oct 22, 2010 8:05am

superluz
| Joined Jul 2008 | Status: Trader | 9 Posts

Wow....thanks a lot 7bit!!

It’s working really nice…

I can’t wait to try some of this packages…
http://robjhyndman.com/category/software/

I’ve just add a back test possibilities to your indicator (if you don’t mind) so that it’s maybe easier to find some optimal parameters.

Thanks again
Claudio

Attached File(s)

AutoregressionBack.MQ4 2 KB | 2,755 downloads

#19
Oct 22, 2010 9:23am Oct 22, 2010 9:23am

scarface9
| Joined Jan 2010 | Status: Trader | 46 Posts

Hi All,

Good day,

Thanks a lot for sharing 7bit. I will try it out soon and see how it goes.

I believe it would be a good expereince for us.

Thanks again.

Best wishes,

SF

#20
Edited 5:25pm Oct 22, 2010 9:43am | Edited 5:25pm

7bit
Joined Mar 2009 | Status: Trader | 1,231 Posts

Quoting noobie1

Disliked

Thanks Steve... got it to open and compile with no errors... got it to open the dialog box on the chart, but it won't go any further than that. Heh... told you I was a noob!!

Ignored

It could be a dozen reasons plus the ones I didn't think of because they never happened to me. Without the debugging output I cannot have any clue what is going on at your site.

Please post the output from DebugView.exe. It should look like the following:

Inserted Code

[37308] installed exception handler for C:ProgrammeBestDirectMT4expertslibrariesmt4R.dll
[37308] <2> TRConsole.Create: trying to start R: C:/Programme/R/R-2.11.1/bin/Rterm.exe --no-save
[37308] <2> TRConsole.SetNotBusy: R is idle
[37308] <2> TRConsole.ExecuteCode: in  >>>  options(digits=15)
[37308] <2> TRConsole.ExecuteCode: out <<<  
[37308] <2> TRConsole.SetNotBusy: R is idle
[37308] <2> TRConsole.Create: R successfully started
[37308] RInit: RHandle = 88A7030 (143290416)
[37308] <2> TRConsole.ExecuteCode: in  >>>  as.integer(as.integer(exists("model")))
[37308] <2> TRConsole.ExecuteCode: out <<<  [1] 0
[37308] <2> TRConsole.SetNotBusy: R is idle
[37308] <2> TRConsole.ExecuteCode: in  >>>  hist <- c(1.3935,1.39245,1.39322,1.39491,1.39111,1.39128,1.38832,1.39004,1.39299,1.39592,1.3961,1.39282,1.39088,1.39064,1.39063,1.39123,1.39158,1.39254,1.39198,1.3923,1.39302,1.39072,1.39383,1.39723,1.40034,1.40046,1.39919,1.402,1.4024,1.40062,1.40211,1.3968,1.39498,1.39157,1.39102,1.39185,1.39213,1.39147,1.39546,1.39763,1.39621,1.39586,1.39642,1.39495,1.39584,1.39687,1.3962,1.39482,1.39678,1.38958,1.3836,1.3852,1.38404,1.38467,1.38201,1.38117,1.37846,1.37673,1.37781,1.37719,1.37687,1.37559,1.37415,1.36987,1.37387,1.37413,1.37253,1.37288,1.37419,1.3777,1.38123,1.3826,1.37904,1.38093,1.38456,1.38717,1.39232,1.39327,1.39086,1.39171,1.38986,1.39052,1.39198,1.39212,1.39359,1.39295,1.39717,1.39628,1.39346,1.39496,1.39361,1.39912,1.39963,1.39838,1.39884,1.39777,1.39669,1.39486,1.39172,1.3889,1.39087,1.38987,1.38638,1.38629,1.38862,1.38759,1.38916,1.38817,1.38808,1.39031,1.39287,1.39512,1.39624,1.39741,1.39757,1.39733,1.39618,1.39647,1.40129,1.39967,1.40103,1.40288,1.40825,1.40955,1.4076,1.40932,1.40971,1.40862,1.4092,1.40534,1.40425,1.40373,1.40296,1.40298,1.40362,1.40254,1.40698,1.40793,1.40828,1.40722,1.40566,1.40731,1.40529,1.40652,1.40845,1.40637,1.4079,1.40842,1.40633,1.40784,1.40748,1.41141,1.40973,1.40776,1.40726,1.40697,1.40791,1.40823,1.40354,1.39828,1.39662,1.39597,1.39613,1.39556,1.39625,1.39634,1.3949,1.39629,1.39592,1.39179,1.39632,1.39633,1.39772,1.39864,1.39726,1.39937,1.39547,1.39633,1.39725,1.39659,1.39707,1.39578,1.39385,1.39201,1.39149,1.39152,1.39242,1.39158,1.39109,1.38541,1.38669,1.38452,1.38328,1.38131,1.38449,1.38227,1.38088,1.38188,1.37892,1.38133,1.38224,1.38369,1.38643,1.38802,1.38773,1.38641,1.38618,1.38652,1.38796,1.38811,1.38775,1.38723,1.38883,1.38849,1.38872,1.38753,1.38941,1.3934,1.39326,1.39425,1.39304,1.3941,1.39443,1.39205,1.39614,1.39669,1.39735,1.39651,1.39754,1.39882,1.39643,1.39958,1.39758,1.40038,1.39357,1.39284,1.3909,1.39033,1.39124,1.39222,1.39324,1.39298,1.38777,1.38824,1.38812,1.3908,1.39087,1.39197,1.3927,1.39314,1.3949,1.39556,1.39567,1.39169,1.39087,1.39243,1.39247,1.39153,1.39255,1.39114,1.39289,1.38988,1.38857,1.39195,1.39401,1.39407,1.40229,1.39881,1.39709,1.39666,1.39634,1.39831,1.39764,1.39236,1.39174,1.39183,1.39096,1.39115,1.39164,1.39154,1.39217,1.39248,1.39299,1.39354,1.39359,1.39317,1.3924,1.39231,1.38907,1.38751,1.38518,1.38604,1.38337,1.38661,1.38588,1.38528,1.38458,1.38411,1.38405,1.38464,1.38291,1.38303,1.38353,1.38534,1.3835,1.38343,1.38377,1.38353,1.38466,1.3856,1.38291,1.3833,1.38382,1.37981,1.38001,1.37683,1.37679,1.37754,1.37758,1.37672,1.37024,1.37011,1.36858,1.36508,1.36568,1.36604,1.36724,1.36479,1.36756,1.36748,1.36862,1.36877,1.36854,1.36888,1.369,1.36865,1.36819,1.37047,1.36868,1.37059,1.36884,1.36841,1.36794,1.37347,1.37614,1.37631,1.37606,1.37602,1.37665,1.37544,1.37528,1.37756,1.37773,1.37882,1.37907,1.37797,1.37775,1.37635,1.37412,1.3753,1.37458,1.37446,1.37377,1.37317,1.37228,1.37382,1.37213,1.36771,1.36487,1.36394,1.36663,1.3659,1.36558,1.36538,1.36498,1.36307,1.36204,1.36286)
[37308] <2> TRConsole.ExecuteCode: out <<<  
[37308] <2> TRConsole.SetNotBusy: R is idle
[37308] <2> TRConsole.ExecuteCode: in  >>>  hist <- c(hist,1.36344,1.3633,1.36397,1.36233,1.36197,1.36003,1.3653,1.36298,1.36764,1.36257,1.36503,1.3662,1.36327,1.36223,1.35877,1.35773,1.36003,1.35953,1.36032,1.36078,1.36303,1.36246,1.36267,1.36218,1.36267,1.36288,1.36404,1.36304,1.36299,1.36354,1.35992,1.35956,1.36037,1.36105,1.36029,1.35938,1.35847,1.36111,1.36136,1.35793,1.35855,1.35857,1.35782,1.35793,1.35773,1.35728,1.35687,1.35848,1.35848,1.35783,1.35651,1.35818,1.35759,1.35813,1.35449,1.34626,1.34669,1.34564,1.34516,1.34779,1.34025,1.3406,1.3404,1.34498,1.34691,1.34598,1.34652,1.34392,1.34428,1.34297,1.34367,1.34398,1.34548,1.34707,1.34738,1.34779,1.34784,1.34744,1.347,1.34567,1.34959,1.34727,1.34637,1.34592,1.34433,1.34594,1.34748,1.34658,1.34696,1.34628,1.34653,1.34578,1.34756,1.34789,1.34777,1.34825,1.34905,1.34914,1.34694,1.34654,1.34721,1.34748,1.34755,1.34646,1.34509,1.34142,1.34114,1.34234,1.33769,1.33126,1.33136,1.33344,1.33338,1.33454,1.33403,1.33272,1.33157,1.32937,1.33178,1.33138,1.33142,1.33176)
[37308] <2> TRConsole.ExecuteCode: out <<<  
[37308] <2> TRConsole.SetNotBusy: R is idle
[37308] <2> TRConsole.ExecuteCode: in  >>>  hist <- rev(hist)
[37308] <2> TRConsole.ExecuteCode: out <<<  
[37308] <2> TRConsole.SetNotBusy: R is idle
[37308] <2> TRConsole.ExecuteCode: in  >>>  ord <- 200
[37308] <2> TRConsole.ExecuteCode: out <<<  
[37308] <2> TRConsole.SetNotBusy: R is idle
[37308] <2> TRConsole.ExecuteCode: in  >>>  meth <- "ols"
[37308] <2> TRConsole.ExecuteCode: out <<<  
[37308] <2> TRConsole.SetNotBusy: R is idle
[37308] <2> TRConsole.ExecuteCodeAsync: starting separate thread
[37308] <2> TRConsole.ExecuteCodeAsync: in  >>>  model <- ar(hist, aic=FALSE, order=ord, method=meth)
[37308] <2> TRConsole.ExecuteCodeAsync: out <<<  
[37308] <2> TRConsole.SetNotBusy: R is idle
[37308] <2> TRConsole.ExecuteCode: in  >>>  as.integer(as.integer(exists("model")))
[37308] <2> TRConsole.ExecuteCode: out <<<  [1] 1
[37308] <2> TRConsole.SetNotBusy: R is idle
[37308] <2> TRConsole.ExecuteCode: in  >>>  ahead <- 20
[37308] <2> TRConsole.ExecuteCode: out <<<  
[37308] <2> TRConsole.SetNotBusy: R is idle
[37308] <2> TRConsole.ExecuteCode: in  >>>  pred <- predict(model, n.ahead=ahead)$pred
[37308] <2> TRConsole.ExecuteCode: out <<<  
[37308] <2> TRConsole.SetNotBusy: R is idle
[37308] <2> TRConsole.ExecuteCode: in  >>>  as.vector(rev(pred))
[37308] <2> TRConsole.ExecuteCode: out <<<   [1] 1.38301336319332 1.38389608665612 1.37922091302647 1.38505219551758 
[37308]  [5] 1.38570987442544 1.39025229696264 1.39402072513305 1.39155695712116 
[37308]  [9] 1.39707648179245 1.39889151642112 1.40033000934452 1.40453666954193 
[37308] [13] 1.40543721177263 1.40936641333182 1.40585754996571 1.40698643957989 
[37308] [17] 1.41049855099731 1.40476938037014 1.40114541852298 1.39789910463703
[37308] <2> TRConsole.SetNotBusy: R is idle
[...]
and so on...

Platform Tech
/
MQL4 -> R-Project - Interface Library
Reply to Thread
- Page 1 2 3 4 5 6 11
- Page 1 2 3 4 11

0 traders viewing now

Options

Similar Threads

MQL4 -> R-Project - Interface Library