BookmarkSubscribeRSS Feed
🔒 This topic is solved and locked. Need further help from the community? Please sign in and ask a new question.
pmdci
Fluorite | Level 6

Hello,

I am using SAS Enterprise miner 4.3 and I have a question about the Scree graph for the DMneural/Pricomp node.

Our lecturer wrote a tutorial explaining that I needed to connect the nodes as follows

INPUT DATA --> DATA PARTITION --> DMNEURAL --> NEURAL NETWORK.

First the exercise asks to run the Neural Network and perform some analysis. So far so good.

However, the exercise then asks for us to open the Dmneural node and look at the Scree Chart. I can't seem to find this chart anywhere! I looked all around. Could someone please advise how to find it?

Thanks in advance for the help.

Regards,

P.

1 ACCEPTED SOLUTION

Accepted Solutions
DougWielenga
SAS Employee

It is not clear what your professor is referring to as the 'Scree' plot.  That term is often used in association with hierarchical cluster analysis and provides a way to assess how many clusters should be retained.  The term scree is used as it refers to the rubble often found at the base of a cliff.  As the number of clusters increases, a great deal more explanation is achieved up to a point (representing a drastic drop in the the error) followed by a smaller more systematic change as additional cluster splits are added (representing the small slope at the bottom or the 'scree').   

 

I would guess that your professor is referring to the Iteration Plot which is generated by the Neural Network node.   It shows great improvement initially followed by lesser improvements to training over time.  If you have a validation data set, the performance on the validation plot will start to get worse.   You can then identify when to stop the iterative process when you have achieved the best values on the validation data and/or on the place where the more change in the performance is much less.   I suspect your professor was referring to the results of the Neural Network node and inadvertently referred to the wrong node since there is no such report generated by the DMNeural node.   

 

I hope this helps!

Doug

View solution in original post

1 REPLY 1
DougWielenga
SAS Employee

It is not clear what your professor is referring to as the 'Scree' plot.  That term is often used in association with hierarchical cluster analysis and provides a way to assess how many clusters should be retained.  The term scree is used as it refers to the rubble often found at the base of a cliff.  As the number of clusters increases, a great deal more explanation is achieved up to a point (representing a drastic drop in the the error) followed by a smaller more systematic change as additional cluster splits are added (representing the small slope at the bottom or the 'scree').   

 

I would guess that your professor is referring to the Iteration Plot which is generated by the Neural Network node.   It shows great improvement initially followed by lesser improvements to training over time.  If you have a validation data set, the performance on the validation plot will start to get worse.   You can then identify when to stop the iterative process when you have achieved the best values on the validation data and/or on the place where the more change in the performance is much less.   I suspect your professor was referring to the results of the Neural Network node and inadvertently referred to the wrong node since there is no such report generated by the DMNeural node.   

 

I hope this helps!

Doug

sas-innovate-2024.png

Join us for SAS Innovate April 16-19 at the Aria in Las Vegas. Bring the team and save big with our group pricing for a limited time only.

Pre-conference courses and tutorials are filling up fast and are always a sellout. Register today to reserve your seat.

 

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
  • 1 reply
  • 869 views
  • 0 likes
  • 2 in conversation