Jump to content

P.NN Ea


Recommended Posts

I want to share with you this predictable NN that i'm working on:

Consist of three files:

Pnn.mqh(experts/libraries/)

PnnTrainer.mq4(/experts/)

PnnSystem.mq4(/experts/)

First you have to train the library with the PnnTrainer ea some especific amount of time and Pair. Depending on the amount of time it could take many hours. Once trained, PNNSystem can trade depeding on the trained data.

This is a beta version and still on development but it is very promising.

If you have any experience in developing, please share your knowlegde!

Enjoy

btw,thanks to pacohernandez whom has the original idea. :peace: :peace: :peace:

[url]http://www.4shared.com/file/94533356/297016e9/predineunet.html[/url]

Link to comment
Share on other sites

Re: P.NN Ea

 

I want to share with you this predictable NN that i'm working on:

Consist of three files:

Pnn.mqh(experts/libraries/)

PnnTrainer.mq4(/experts/)

PnnSystem.mq4(/experts/)

First you have to train the library with the PnnTrainer ea some especific amount of time and Pair. Depending on the amount of time it could take many hours. Once trained, PNNSystem can trade depeding on the trained data.

This is a beta version and still on development but it is very promising.

If you have any experience in developing, please share your knowlegde!

Enjoy

btw,thanks to pacohernandez whom has the original idea. :peace: :peace: :peace:

[url]http://www.4shared.com/file/94533356/297016e9/predineunet.html[/url]

 

Thanks very much for sharing this EA based on a different type of Neural Networks. A special thank and think go to Paco hernandez for his interesting idea.

 

The programmer of Dhananjaya EA and PriceTypeA EA have assembled the three separate files in a one file, its name is pnn_system_demo (you can find it as attachment). The result is an EA that don't need to be trained by the user. the EA do it.

 

The importance of introducing pnn (Probabilistic Neural Networks) is to get a self learning EA, it means that your EA will continuously learn from the past trades and automatically modify the future trades entry or exit.

The use of pnn in trading is in the begining stage, exept this system i haven't find any other EA that uses Pnn. I think that any EA programmer needs a such thing in order to get more and more profit.

 

any one have Dhananjaya EA and PriceTypeA EA? it seems to be more interesting than pnn_system_demo :)

 

 

Regards

Mourad

Link to comment
Share on other sites

Re: P.NN Ea

 

Thanks for your contribution!

I have reviewed the code and it is also needed a training with the past data but the difference is that once trained, the ea train itself ahead.

I think that Dhananjaya EA and PriceTypeA EA, may use a very similar model to approach trading.

If anybody has it please share..! :arrow: :arrow:

Link to comment
Share on other sites

Re: P.NN Ea

 

Thanks for your contribution!

I have reviewed the code and it is also needed a training with the past data but the difference is that once trained, the ea train itself ahead.

I think that Dhananjaya EA and PriceTypeA EA, may use a very similar model to approach trading.

If anybody has it please share..! :arrow: :arrow:

 

Hi Daw-Veed :)

 

I don't think that EA need a first training. It just do it itself: look the line 63 of the code where the EA uses the function StartTraining1(int shift) for training, and in the line 72 the EA uses the function PNN1Save() to store the training results in the pnn1.dat file. the file pnn1.dat will be called by PNN1Load() in order to be used for classifying.

 

I have the same request too:

 

"Please share Dhananjaya EA and PriceTypeA EA!! :? "

 

regards

Mourad

Link to comment
Share on other sites

Re: P.NN Ea

 

Hi Daw-Veed :)

 

I don't think that EA need a first training. It just do it itself: look the line 63 of the code where the EA uses the function StartTraining1(int shift) for training, and in the line 72 the EA uses the function PNN1Save() to store the training results in the pnn1.dat file. the file pnn1.dat will be called by PNN1Load() in order to be used for classifying.

 

I have the same request too:

 

"Please share Dhananjaya EA and PriceTypeA EA!! :? "

 

regards

Mourad

 

Hi Mourad,

Yes, it's right. The ea train itself the number of bars specified in TestRange input. The issue is that you dont know when does it finish to train past data.. I think it could be fixed by a warning..

Link to comment
Share on other sites

Re: P.NN Ea

 

Nice share. Thanks.

 

But, do you guys have any backtest results? Could you attach the graph? Thanks.

 

Regards,

Ore no Shinka Hikari yo Hayai. Zen Uchi o Nani no Mono Ore no Shinka Chuito Kore Nai.

Ten no Michi yo Iki. Subete o Sukosadoru Otoko.

Link to comment
Share on other sites

Re: P.NN Ea

 

Nice share. Thanks.

 

But, do you guys have any backtest results? Could you attach the graph? Thanks.

 

Regards,

 

Hi William,

You will find as attachment a backtest of the pnn_system_demo for the period between january 12 and march 19 of this year.

 

:)

Link to comment
Share on other sites

Re: P.NN Ea

 

:)

 

I have a remark about this EA.

 

a Probabilistic Neural Network (Pnn) is an algorithm that allow us to classify past data in very short time comparing with other algorithms.

 

For This EA, the pnn uses a very simple strategy. We can introduce a pnn into any other EA that uses a different strategy. The result is an EA that continuously learn and improve the strategy.

 

In practice, every thing is in a begining stage. Me too I am trying to improve an EA using a pnn. But a lot of work must be done before getting some results. But one thing is sure, any results obtained with pnn will be great.

 

regards

:)

Link to comment
Share on other sites

Re: P.NN Ea

 

Hi Mourad

I'm going to start forward-testing in vps . If you have a more suitable ea (like Dhananjay ea :) ) please let me know..

 

Regards

 

Hi Daw-Veed :)

 

Thank you for the VPS. But the EA that i am now working on is based on the black dog strategy (you can find it on this wonderful forum), and i am using a pnn to improve the results.

When i finish the EA i will surely share it here.

 

For Dhananjay ea, me too i am looking for but unfortunately :cry:

 

Regards

:)

Link to comment
Share on other sites

  • 2 months later...
  • 2 weeks later...

Re: P.NN Ea

 

I have been looking at these PNN (Probabilistic Neural Networks) methods this week. I found the hernandezgomez code from the original site

 

http://www.hernandezgomez.com/index.php/using-a-pnn-probabilistic-neural-network-and-metatrader-4-mql4-to-trade-forex/

 

My own feeling is that this code is unfortunately "crap". There needs to be better "classifiers" in the code for it to work.

 

I am still trying to figure out how to make a PNN to work, if anyone has a clue I would appreciate a frank explanation.

 

By the way, if you search for successful EAs, you will quickly discover that correctly constructed PNNs statistically dominate (outperform) most other strategies. So this topic is important.

Link to comment
Share on other sites

Re: P.NN Ea

 

Hi,

hernandezgomez made a misstake in calculate the pdf for pnn.

for a nice pnn and code it in c++, please referer to http://codebase.mql4.com/ru/5704

I translated this page by google and my impression is that it is a regular back propagation NN

so it is different from Hernandez' PNN, can you provide good description in plain english for

PNN to back your claim?

I would like to learn about PNN but so far I cant find good and easy reading material :(

from what I red so far PNN is kind of easy to implement in MQL compare to back propagation

(that why it is implemented in dll)

 

Regs,

Lego

Link to comment
Share on other sites

  • 11 months later...

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...