Discussion of article "Programming a Deep Neural Network from Scratch using MQL Language" - page 3

 
Roman Shiredchenko #:

Is there any way to incorporate error back propagation into this network?

For some reason there is an opinion that such a network would be more flexible and provide better inputs - outputs....

A network with error back propagation is not a completely different network, is it?

See here - https://www.mql5.com/en/code/27396 for the error back propagation network in MQL5.

This is a transposition of the original Sish code from here - https://www.mql5.com/en/code/9002 - there is a theory there in brief.

BPNN MQL Predictor Demo with library
BPNN MQL Predictor Demo with library
  • www.mql5.com
This is a demo indicator with BPNN neural network library ported from C++ to MQL.
 
Stanislav Korotky #:

See here - https://www.mql5.com/en/code/27396.

This is a transfer of the original Sish code from here - https://www.mql5.com/en/code/9002 - there is a theory there in brief.

I'll have a look...

 

Hi Andy:

Congratulations on your article and your program. I just tested it in MT5 and it seems to work correctly, although I think that the neural network is not able to generalize and outside the backtest period the results worsen significantly.

Thank you very much for your work and for the excellent idea of leaving the training part to the strategy simulator.

Regards

Gerardo

 

A rare and good article, thanks for sharing.

Deep learning is indeed a good way to do trading, the only problem is the impact caused by anomalous data

The age of AI is here!

 
The animation is super, the author has gone to a lot of trouble
 

Hello, your article was exceptional, very intelligent and attractive, but when I went to compile it to test it on my account it showed errors. could you check what happened? I am very anxious to test the EA of your article. Thank you very much.



 
Hi Sir good day i was indeed fascinated by your EA regarding on Neural Network and
I was working on it for quite a while now
I dont know whats wrong on my code since i'm not an expert i'm just an IT student studying in a community college
I would like to make your EA into custom indicator that just make object label
(Buy [+1], Sell[-1],ignore[0] also sir i change rates[0].high to rates[1].high) but no matter how i read some documentation it doesnt show the final output
 
Karlis Balcers #:

Update the following function to return `bool` instead of `void` and you will see that there was a bad number of weights given.

Note that you also need to update the wheights on top of the file (it's not enough to update them only when you initialize the network :P 

It's a hard-coded NN, I don't recommend increasing the number of neurons as you won't be able to train it later...

 
Very nice article, thanks for this great introduction. When I run the code I allways get 0.33333331 as yValues and therefore no trade. I checked with a bunch of instruments. Any idea what I'm doing wrong?
 
Thanks for the codes you shared. As an new coder, I have some questions to ask, hopefully can get some helps from your experience guys. Thanks in advance.

1. When I compile the file "DeepNeuralNetwork.mqh", I have some errors and couldn't fix them.

2. In DemoEADNN.mq5, couldn't find this two files, see the screenshot below.

Thanks.

Files: