[prev in list] [next in list] [prev in thread] [next in thread] 

List:       wekalist
Subject:    [Wekalist] Re: Neural Network soft output
From:       Eibe Frank <eibe () cs ! waikato ! ac ! nz>
Date:       2003-12-21 20:46:05
[Download RAW message or body]

Hi Dave,

I think you can only get that info in the Explorer (by selecting the 
"Output text predictions on test set" option under "More options...").

Cheers,
Eibe

On Friday, December 19, 2003, at 12:57  PM, 
wekalist-request@list.scms.waikato.ac.nz wrote:

> From: "Dave Bazell" <bazell@comcast.net>
> Date: Thu Dec 18, 2003  5:37:34  PM Pacific/Auckland
> To: <wekalist@list.scms.waikato.ac.nz>
> Subject: [Wekalist] Re: Neural Network soft output (Eibe Frank)
> Reply-To: Dave Bazell <bazell@comcast.net>
>
>
> Hi Eibe,
>
> thanks for the information.   I am using the command line.  I have used
> the -p option on the command line but it appears to only give me the
> activation of the predicted output class.  What I am looking for is the
> activation of all the classes.  Is that available?
>
> For example, when I give the -p 1-13 option, I get the following:
>
> 5110 2 0.9875484829303337 2
> (1.07,1.97,1.55,26.62,21.08,0.2,0,3.96,2.79,1.79,21.37,21.85,20.79)
>
> In interpret this as the 5110th example, true class 2; probability of 
> class
> membership 0.9875...; predicted class 2; input parameters.  I would 
> like the
> probability of class membership for all the classes.
>
> Thanks,
>
> Dave


_______________________________________________
Wekalist mailing list
Wekalist@list.scms.waikato.ac.nz
https://list.scms.waikato.ac.nz/mailman/listinfo/wekalist
[prev in list] [next in list] [prev in thread] [next in thread]