Date:         Tue, 25 Jul 2000 15:45:16 +0200
Reply-To:     Claus Gotfred Rasmussen <CGR@ACCEPTCARD.DK>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Claus Gotfred Rasmussen <CGR@ACCEPTCARD.DK>
Subject:      Re: What's the Output Dataset?
Comments: cc: "Kattamuri.Sarma@RESPONSEINSURANCE.COM"
          <Kattamuri.Sarma@RESPONSEINSURANCE.COM>,
          "Mark.K.Moran@CCMAIL.CENSUS.GOV" <Mark.K.Moran@CCMAIL.CENSUS.GOV>
Content-Type: text/plain; charset="iso-8859-1"

Hi,

Here is another opportunity for me to advertise for the specialized SAS/EM
mailing list located at eGroups.

Addresses:
Post message: saseminer@egroups.com
Subscribe:  saseminer-subscribe@egroups.com
Unsubscribe:  saseminer-unsubscribe@egroups.com
List owner:  saseminer-owner@egroups.com
URL: http://www.egroups.com/group/saseminer


Regarding the output dataset, I think the method described by Kattamuri is
the only way.

Kind regards,
Claus

-----Original Message-----
From: Kattamuri.Sarma@RESPONSEINSURANCE.COM
[mailto:Kattamuri.Sarma@RESPONSEINSURANCE.COM]
Sent: Tuesday, July 25, 2000 3:30 PM
To: SAS-L@LISTSERV.UGA.EDU
Subject: Re: What's the Output Dataset?


Mark:
   Let me describe how I found the output data set created by the score
node. I hope this
   example will help you:

   My flow diagram looks something like the following:

     Raw data (input node) ---> Neural network node |

      input data set (input node) ---------------------- >     Score node
---> SASCode node

     I could not draw the diagram well, but there should be an arrow from
Neural Network node to score node.
      From raw or training dat sets is first connected to neural network
node.  I picked another input  node,
      connected it to score node. (make sure the role is set to score, in
the input data set),  I also connected
     the neural network node to the score node, so that the score node has
access to to the neural model.
     Then from score nod I drew an arrow to the SAS code node.

      If you open the SAS code node,  press the macros tab, and presse the
data sets / all   undertab ,
      you find the name of the scored data set.   I used that data make
some comparisons
      of  predicted (P_target ) and target variables.   It all worked fine.

        I hope this works for you also, and I am posting for the list , so
that if someone has a better solution,
     I would like to hear.

     Kattamuri