AForge.NET

  :: AForge.NET Framework :: Articles :: Forums ::

Regularization or Dropout

The forum is to discuss topics from different artificial intelligence areas, like neural networks, genetic algorithms, machine learning, etc.

Regularization or Dropout

Postby Lawrence » Mon Jun 06, 2016 2:01 pm

Hi,

I am trying to cut down on the fitting in my net.

Is the ability to use a Regularization parameter (adding a term for every weight in the net) built-in somewhere?

Could you please point me at any documentation.

OR perhaps a Dropout approach is available?
Lawrence
 
Posts: 24
Joined: Tue May 17, 2016 11:32 am

Re: Regularization or Dropout

Postby andrew.kirillov » Mon Jun 06, 2016 5:16 pm

Hello,

Lawrence wrote:Could you please point me at any documentation.

Here it is.

Non of the mentioned features are implemented, since not mentioned explicitly in documentation.
With best regards,
Andrew


Interested in supporting AForge.NET Framework?
User avatar
andrew.kirillov
Site Admin, AForge.NET Developer
 
Posts: 3453
Joined: Fri Jan 23, 2009 9:12 am
Location: UK

Re: Regularization or Dropout

Postby Lawrence » Tue Jun 07, 2016 8:50 am

Andrew,

Thank you for documentation head.

As you have not implemented Regularization in your excellent library, do you favour an alternative approach to achieving the goal of mitigating over-fitting? If so could you point me in that direction please.
Lawrence
 
Posts: 24
Joined: Tue May 17, 2016 11:32 am




Return to Artificial Intelligence