BookmarkSubscribeRSS Feed
SlutskyFan
Obsidian | Level 7
Is is possible to call for an actual plot/network visualization of the network constructed using the nerual network node in EM?

( I know for very complicated networks with lots of hidden layers it could be too cluttered to interpret but I would still like this option)
2 REPLIES 2
WayneThompson
SAS Employee
An interactive neural network buidler to define and visualize the network is a feature we would like to add to EM. thanks
SlutskyFan
Obsidian | Level 7
Thanks- if that would be similar to the interactive decision tree function that would be great.

I did find a NESUG 2007 paper :

http://www.nesug.org/proceedings/nesug07/po/po23.pdf

that mentioned extracting output from the EM neural network results, and uses a macro to produce a 'constellation diagram' that supposedly does this. I did extract the data I thought I would need, but don't know how to configure it to work in the macro they provided. (+ it was likely related to an older version of EM)

This reminds me roughly of what you would do in relation to social network analysis, where you would essentially construct a contiguity matrix (0's and 1's- an edge list? ) that indicate # links between nodes to construct a network. (With something like UCINET )

I know it is not essential to actual model deployment, but it would be a nice visualization tool to show management an example of a simple network and how it works and my own curiousity (in the mean time prior to the next version of EM). Although I'd rather figure out how to do this in SAS vs. rigging something in UCINET if at all possible).

Any suggestions on the macro? Would this be something SAS tech support could help me with if I called or put in a request?

sas-innovate-2024.png

Available on demand!

Missed SAS Innovate Las Vegas? Watch all the action for free! View the keynotes, general sessions and 22 breakouts on demand.

 

Register now!

How to choose a machine learning algorithm

Use this tutorial as a handy guide to weigh the pros and cons of these commonly used machine learning algorithms.

Find more tutorials on the SAS Users YouTube channel.

Discussion stats
  • 2 replies
  • 874 views
  • 0 likes
  • 2 in conversation