Gamestudio Links
Zorro Links
Newest Posts
Data from CSV not parsed correctly
by jcl. 04/26/24 11:18
M1 Oversampling
by jcl. 04/26/24 11:12
Why Zorro supports up to 72 cores?
by jcl. 04/26/24 11:09
Eigenwerbung
by jcl. 04/26/24 11:08
MT5 bridge not working on MT5 v. 5 build 4160
by EternallyCurious. 04/25/24 20:49
Trading Journey
by howardR. 04/24/24 20:04
Zorro FIX plugin - Experimental
by flink. 04/21/24 07:12
Scripts not found
by juergen_wue. 04/20/24 18:51
AUM Magazine
Latest Screens
The Bible Game
A psychological thriller game
SHADOW (2014)
DEAD TASTE
Who's Online Now
3 registered members (AndrewAMD, wandaluciaia, 1 invisible), 798 guests, and 6 spiders.
Key: Admin, Global Mod, Mod
Newest Members
wandaluciaia, Mega_Rod, EternallyCurious, howardR, 11honza11
19049 Registered Users
Previous Thread
Next Thread
Print Thread
Rate Thread
I am insisting with Machine Learning #463932
01/07/17 21:43
01/07/17 21:43
Joined: Jul 2016
Posts: 64
G
gtell Offline OP
Junior Member
gtell  Offline OP
Junior Member
G

Joined: Jul 2016
Posts: 64
But I still do not understand some things.
This post is a question to whom has more experience of me.
The question is very simple and regards how Zorro passes data to the R script.
To put in short, consider the following robot:

Code:
#include <r.h>

var change(int n)
{
	return scale((priceClose(0) - priceClose(n))/priceClose(0),100)/100;
}

var range(int n)
{
	return scale((HH(n) - LL(n))/priceClose(0),100)/100;
}


function run()
{
	StartDate = 2012;
	BarPeriod = 60*24;
	LookBack = 100;

	WFOPeriod = 700; // 1 year
	DataSplit = 90;
	NumCores = -1;  // use all CPU cores but one

	set(RULES);
	set(LOGFILE);
	Spread = RollLong = RollShort = Commission = Slippage = 0;
	//LifeTime = 3;
	int Offset = 0;
	if(Train) {
	   Hedge = 2;
	   Offset = 1;
   }
	
	
	var c1=change(1+Offset), c2=change(2+Offset), c3=change(3+Offset), c4=change(4+Offset);
	var r1=range(1+Offset), r2=range(2+Offset), r3=range(3+Offset), r4=range(4+Offset);
	
	var pred = adviseLong(NEURAL,change(1),c1,c2,c3,c4,r1,r2,r3,r4);
	if(Train) enterLong();  
	
	if(Test) {
		print(TO_LOG, "nSignals: %1.3f,%1.3f,%1.3f,%1.3f,%1.3f,%1.3f,%1.3f,%1.3f", c1,c2,c3,c4,r1,r2,r3,r4); 
		print(TO_LOG, "nPrediction value: %1.3f", pred);
	
		if(pred>0.5 && NumOpenLong<1) enterLong();  
   	else if(pred<0.5 && NumOpenShort<1)  enterShort(); 
   }
}



Let's not consider the efficiency of this startegy.
But, let's have a look to the log:

Code:
[1024: Wed 16.12.15 00:00] -160 -4 54/96  (1.09902)
Signals: -0.298,-0.186,-0.004,-0.149,0.400,0.081,-0.104,-0.244
Prediction value: 0.252
[EUR/USD::L2250] Reverse 1@1.09341: -4.43 at 00:00:00
[EUR/USD::S2451] Short 1@1.09341  at 00:00:00

[1025: Thu 17.12.15 00:00] -164 +5 55/96  (1.09329)
Signals: -0.320,-0.405,-0.250,-0.129,0.331,0.273,0.094,-0.045
Prediction value: 0.237

[1026: Fri 18.12.15 00:00] -164 +8 55/96  (1.08767)
Signals: -0.177,-0.326,-0.340,-0.261,-0.099,0.347,0.405,0.253
Prediction value: 0.428

[1027: Mon 21.12.15 00:00] -164 +6 55/96  (1.08404)
Signals: 0.130,-0.022,-0.139,-0.218,-0.199,-0.275,0.173,0.238
Prediction value: 0.739
[EUR/USD::S2451] Reverse 1@1.08695: +5.63 at 00:00:00
[EUR/USD::L2752] Long 1@1.08695  at 00:00:00

[1028: Tue 22.12.15 00:00] -158 +4 56/96  (1.08671)
Signals: 0.183,0.224,0.129,-0.019,-0.054,-0.022,-0.191,0.027
Prediction value: 0.698

[1029: Wed 23.12.15 00:00] -158 +7 56/96  (1.09106)
Signals: 0.168,0.252,0.297,0.194,-0.116,-0.016,0.022,-0.082
Prediction value: 0.735



So, what I expect is that the signals were shifted, if in one bar I get:

0.1, 0.5, -0.2, -0,3

the next bar I should have something like:

0.3, 0.1, 0.5, -0.2

the first value is new and the other are taken from the previous bar.
I do not understand why this is not happening.
Is someone able to explain me?

Thanks.
Cheers

Re: I am insisting with Machine Learning [Re: gtell] #463950
01/09/17 11:22
01/09/17 11:22
Joined: Jul 2016
Posts: 64
G
gtell Offline OP
Junior Member
gtell  Offline OP
Junior Member
G

Joined: Jul 2016
Posts: 64
Sorry, I have got it.
change gives values are not time invariant.


Moderated by  Petra 

Powered by UBB.threads™ PHP Forum Software 7.7.1