We’re smarter together. Learn from this collection of community knowledge and add your expertise.

SAS Visual Statistics: Getting Started - Ask the Expert Q&A

by SAS Employee MelodieRush ‎04-17-2017 05:02 PM - edited ‎07-26-2017 02:54 PM (1,607 Views)

Did you miss the Ask the Expert session on SAS Visual Statistics: Getting Started? Not to worry, you can catch it on-demand at your leisure. Slides from the session included on this post.
 
Watch the webinar
 
The session covers Getting Started with SAS Visual Statistics. You learn how to:

    • Interactively create and refine descriptive and predictive models – for insights fast and bright as lightning.
    • If you’re a business analyst, statistician or data scientist who explores data or builds models, this webinar is for you.
    • Learn about SAS Visual Statistics and its capabilities (including its seamless integration with SAS Visual Analytics), including how to:

                      * Apply statistical modeling techniques to explore data visually.

 

                      * Create models, including decision trees and logistic regression.

 

                      * Compare models by looking at different criteria

 

Here are some highlighted questions from the Q&A segment held at the end of the session for ease of reference.


Q: ­Do we need SAS Visual Analytics (VA) to use SAS Visual Statistics (VS)?­
A: Yes, you need SAS Visual Analytics to use SAS Visual Statistics. SAS Visual Statistics is an add-on product to SAS Visual Analytics.


Q: In the decision tree, can you manually choose the 1st node to split on other than age?
A: Yes, you can build interactive decision trees by right mouse clicking on any node and selecting prune to cut back any growth below that node, Split to select which variable to split on, Split best to let SAS Visual Statistics select the best variable to split on or Train to train out the tree from that node.­


Q: Is it possible to download this data?
A: This data is available as part of the try before you buy and to sign up to use it click here. This data is pretty large, downloading it may not be feasible.


Q: What Data Imputation method is used?
A: For the informative missingness option measure variables are replaced with observed mean and category variables are considered a distinct level.


Q: What variable selection method is chosen?
A: For Linear and Logistic regression when variable selection is checked SAS Visual Statistics uses backward selection on the input variables to determine the most significant variables.


Q: How to export Models built using SAS Visual Statistics to SAS Enterprise Miner, so you could refine them using it, is that possible?
A: You can use the Model Import Node in SAS Enterprise Miner in combination with the Score code created in SAS Visual Statistics. This will allow you to compare the models created in SAS Visual Statistics with models create in SAS Enterprise Miner. You will not be able to tweak the model from SAS Visual Statistics in SAS Enterprise Miner, but you will be able to compare them.


Q: Where does SAS Visual Statistics compare to SAS Enterprise Miner? 
A: SAS Visual Statistics is an in-depth GUI driven approachable modeling solution. It provides quick point and click access to some common modeling techniques and makes these techniques available to users with some statistical skills. SAS Enterprise Miner includes many robust production modeling tools that provide for easy repeatability and easy operationalization. Visual Statistics is great for rapid modeling and prototyping. You can create great models quickly. But for production modeling after exploratory modeling Enterprise Miner provide a robust environment for Sampling, Modifying, and includes many other modeling and machine learning methods. Here's a link with more detail information within our SAS Communities

 

Q: Can you see adjusted R2 instead of the standard R2? Adding more variables will always increase the standard R2 because it doesn't account for the number of variables in the model
A: Yes, Adjusted R2 is an option in Visual Statistics. At the top of the analysis you click on R-Square and a drop down list of options is available. For Logistic Regression the list includes AIC, AICC, BIC, Max-rescaled R-Square, R-Square and -2 Log Likelihood.


Want more tips? Be sure to subscribe to the Data Mining Library to receive follow up Q/A, slides and other related resources from the webinar. From the Data Mining Library, just click Subscribe from the orange bar underneath the list of the recent articles.

Contributors
Your turn
Sign In!

Want to write an article? Sign in with your profile.