Lets build a neural network based scalper together

[jjesteban]

Hey neuro, I had a look at your source code and it doesn't look like you're normalizing the input data. It looks like you're directly feeding bar close prices into the network. Neural network data has to be normalized either from -1 to 1 or 0 to 1.
Einverstanden. Wenn Sie mit Daten in einem anderen Bereich als [-1; 1] oder [0; 1] arbeiten, müssen Sie diese Daten normalisieren. Hier ist die Theorie: - Sie bestimmen den Datenbereich. Angenommen, Sie arbeiten an EURUSD, sodass Ihr Datenbereich etwa [1.1; 1.6] beträgt. - Sie konvertieren alle Daten in den Bereich [0; 1] (oder [-1; 1], falls erforderlich). - Sie verarbeiten die Daten durch die ANN - Sie verkleinern die resultierenden Ausgaben mit derselben ursprünglichen Skalierung. In diesem Fall konvertieren Sie alle Daten aus dem Bereich [0; 1] zurück in [1.1; 1.6]. Hier ist eine in MQL4 portierte Funktion von Atmel-C (C-basierte Sprache zum Programmieren von Atmel AVR-Prozessoren), die genau das tut: Eingefügte Code-Doppelzuordnung (doppeltes x, doppeltes in_min, doppeltes in_max, doppeltes out_min, doppeltes out_max) {return ((x - in_min) * (out_max - out_min)(in_max - in_min) out_min); } Beispiel: So konvertieren Sie x vom Bereich [0,5; 2,5] in den Bereich [0; 1]: Eingefügter Code x2 = map (x, 0,5,2,5,0,1);

[Aipkrmenta74]

Gezeichnet! Ein Tester zur Liste hinzugefügt!

[adrijtte]

Eigentlich verstehe ich, dass * alle * NNs normalisierte Eingänge erfordern. Abgesehen von diesem Problem habe ich mir den Code noch einmal angesehen und er * normalisiert * die Preisdaten auf den Bereich von 0 bis 1. Es war auf den ersten Blick einfach nicht offensichtlich.

[eugenioborregogarcay]

this it not compulsory with FANN. it is highly recommended though.

[adrijtte]

Hey neuro, I had a look at your source code and it doesn't look like you're normalizing the input data. It looks like you're directly feeding bar close prices into the network. Neural network data has to be normalized either from -1 to 1 or 0 to 1.

[uribernal]

Can you pls explain how to install these files... Do I need to install the files in this post only or do I need install the previous files also....
You need the dlls from the previous post. Onu - that account has around 10 different EAs running so you cant judge it by neuraltraders accounts performance - to look at prophetFx - check this account
http://www.mt4i.com/users/mohan2

[Kush]

Folks this new version has been generating pips very consistently I also tried it on USDCHF H4 and pretty good so far - take it for a test drive and let me know.
Führt es so aus?
https://www.tradingintuitive.com/attachments/1529197701.png
http://www.myfxbook.com/members/neuraltraders/neuraltraders/19319-----------------------------------------
http://www.myfxbook.com/members/pc8multifx

[turrete777]

OK First Step - Install the fann2mql package from here
http://go2.wordpress.com/?id=725X1342site=fann2mql.wordpress.comurl=http://www.pipscomfort.com/Fann2MQL/Fann2MQL%200.1.2.msi Then copy the attached NextClose-V13.ex4 file into your experts folder and restart MetaTrader, if you are still getting an error let me know - I have been testing it on GoMarkets Mt4 which you can get from here neuraltraders.info/robots-org/gt4setup.exe Also attached is the updated source code - tulip - will start on documenting and cleaning...
Können Sie bitte erklären, wie diese Dateien installiert werden? Muss ich nur die Dateien in diesem Beitrag installieren oder muss ich auch die vorherigen Dateien installieren?

[uribernal]

Leute, diese neue Version hat sehr konsequent Pips generiert. Ich habe sie auch auf USDCHF H4 ausprobiert und war bisher ziemlich gut. Machen Sie eine Probefahrt und lassen Sie es mich wissen.

[uribernal]

OK First Step - Install the fann2mql package from here
http://go2.wordpress.com/?id=725X1342site=fann2mql.wordpress.comurl=http://www.pipscomfort.com/Fann2MQL/Fann2MQL%200.1.2.msi Then copy the attached NextClose-V13.ex4 file into your experts folder and restart MetaTrader, if you are still getting an error let me know - I have been testing it on GoMarkets Mt4 which you can get from here neuraltraders.info/robots-org/gt4setup.exe Also attached is the updated source code - tulip - will start on documenting and cleaning up the code - btw you can have a look at this article to get the basics : articles.mql4.com/777
https://www.tradingintuitive.com/attachments/1529197712787853704.ex4
https://www.tradingintuitive.com/attachments/15291977141511899796.mq4

[Miraym23f]

hi, i check the file... actually, i don't even know where to put all of those file. which one is expert or indior, and the other file, what is that? are those expert or indicaator?

[Mimimi1928]

I'd like to help as I was just thinking about doing some NN stuff for MetaTrader myself. I think the first thing you need to do neuro if we're really going to make this a collaborative project is to go through and clean up the source code and comment everything properly. A lot of the variable names are very vague and the code is strewn all about the file so jumping back and forth to trace everything is a pain, especially for someone like me who doesn't have experience using these NN libraries. Do you think you could go through and comment all the code thoroughly and clean it up to look nicer so we can all start from a good codebase? If we're going to develop something to be open source we have to make sure good coding standards are followed from the beginning.

[Kush]

OK. thx. Still this thing doesn't work..error 126

[uribernal]

Here is the updated zip - added FannDoubleMT.dll (forgot to add in the previous post) , also all copyright and proprietary stuff has been removed.
https://www.tradingintuitive.com/attachments/15291977081646716661.zip

[uribernal]

Onu, Yes I am on the Neuraltraders staff - but this particular EA - NextClose will always be in the open source domain, maybe we can also rope in a few good people from sourceforge etc. Collaborative effort develops everyone - even companies like sun oracle develop software with open source contributions - so copyrights are not an issue, anyways I will remove all neuraltrader related stuff. 7Bit, Thanks for your feedback, I think we should benchmark with both 2 3 hidden layers and may be look at cascade trading to come up with optimal number of ns but I think we all will agree rather than making it versatile (multi pair) lets focus on one pair and EURCHF with its relatively low spread and rangebound nature will be a good candidate - but I am open to suggestions

[Gadi87]

I however had to modify Fann2mql source code to allow for 3 hidden layers (by default it allows for only 2 ). In another project on SP futures we found 3 layers produced smaller mse than 2 though it took longer to train
IIRC Sie werden nie mehr als 2 versteckte Ebenen benötigen, da 2 Ebenen bereits ausreichen, um eine Funktion zu approximieren, die Sie mit 3 oder mehr Ebenen approximieren könnten. Ich kann mich nicht genau erinnern, wo ich das gelesen habe, aber IIRC hat dies irgendwie mathematisch bewiesen.

[Kush]

The copyright of the EA belongs to NeuralTraders.info U are the guy behind this membership thing? Come on, we ain't here to develop your EAs, if this is the case, or to violate other peoples copyrights!
http://www.myfxbook.com/community/success-stories/explosive-growth-through-ai-/19892,1Jemand scheint auf einer Promotion-Tour zu sein .....
https://www.tradingintuitive.com/attachments/1529197701.png ---------------------------------------------
http://www.myfxbook.com/members/pc8multifx

[uribernal]

Thanks Ronald for your interest and Onu and tulipeverige for your offer to test. I am not a big fan of reinventing the wheel so chose to use the excellent and very mature FANN library, then somebody did the grunt work of importing it into mql - Fann2mql project, I however had to modify Fann2mql source code to allow for 3 hidden layers (by default it allows for only 2 ). In another project on SP futures we found 3 layers produced smaller mse than 2 though it took longer to train - more than 3 layers didn't improve mse by much. Instructions: Copy the dlls into experts/libraries , ntann and ntcommons into includes.
https://www.tradingintuitive.com/attachments/15291977041570302676.zip

[rubenciorr]

hi neuro, count me in! please post more details, that we can start with the developement thanks, -ba

[Parcero]

Thats great.... I can do program little and test a lot..

[Miraym23f]

hi.... i will testing it... i am waiting this one to be develop..

[Kush]

I'm interested to test it ---------------------------------------------
http://www.myfxbook.com/members/pc8multifx

[CarlttGC512]

What form of NN are you using?

[uribernal]

I already have some working code that has been squeezing out some pips on ERUCHF - but is rough around the edges, I believe we have some geniuses here and if we put our heads together we can create something decent (I for one need it to get out of my 9-5 rut) . I have a masters in computer science and have been trading forex, futures options for quite some time.

If this thread invokes some response, I will post my code and we can build up on it. Also if non programmers are interested in testing the system - feel free to post, the EA is visual so you can very well use it for manual scalping too.