BookmarkSubscribeRSS Feed
🔒 This topic is solved and locked. Need further help from the community? Please sign in and ask a new question.
b_smsha
Obsidian | Level 7

Hi, 

 

I keep getting an error when I try to run a multiplot or impute node for my dataset. It gives me this following error https://support.sas.com/kb/20/054.html.

 

After reading the solution, I believe my attributes have proper roles and Ive posted a picture below. 

I have approximately 6051 records and I know that the default level is 512. Can someone help me understand that if I increase this level, what would be the appropriate amount and how would I know or judge this ? 

b_smsha_0-1622667389825.png

armed has various categories.

body cam is 1 or 0.

County, City and State are all nominal data with all the areas in US.

Date has been preprocessed into day, month and year. 

geocoding is 1 or 0

manner of death is shot or tasered. 

mental illness is 1 or 0.

VAR1 is the id. 

 

The dataset is the one from washington post. 

https://github.com/washingtonpost/data-police-shootings

 

Any changes ive made to the data are listed above such as 1,0 for mental ilness etc. 

Please help me out, i've been struggling greatly this whole day.

 

Thank you so much!

 

1 ACCEPTED SOLUTION

Accepted Solutions
MelodieRush
SAS Employee

Hello, This should be a simple solution. The 512 max limit is referring to the number of levels you have for nominal or ordinal variables. I took a look at the original data and it looks like city and name both have over 128 and these are probably your culprits. Try rejecting these (ie not using them in your analysis) and see if you can get the nodes to run. 

 

MelodieRush_0-1622668701119.png

 

If indeed you want to use these in your analysis you will need to up the max levels to the highest number of levels you have. You can do this in the Project Macro Variables window (see below). Keep in mind that by doing this it may increase the time it takes for models to run. For example if city has 1000 unique values when you run a regression it will create 999 parameters for city.

MelodieRush_1-1622668927915.png

 

 

Catch the SAS Global Forum keynotes, announcements, and tech content!
sasglobalforum.com | #SASGF



View solution in original post

3 REPLIES 3
MelodieRush
SAS Employee

Hello, This should be a simple solution. The 512 max limit is referring to the number of levels you have for nominal or ordinal variables. I took a look at the original data and it looks like city and name both have over 128 and these are probably your culprits. Try rejecting these (ie not using them in your analysis) and see if you can get the nodes to run. 

 

MelodieRush_0-1622668701119.png

 

If indeed you want to use these in your analysis you will need to up the max levels to the highest number of levels you have. You can do this in the Project Macro Variables window (see below). Keep in mind that by doing this it may increase the time it takes for models to run. For example if city has 1000 unique values when you run a regression it will create 999 parameters for city.

MelodieRush_1-1622668927915.png

 

 

Catch the SAS Global Forum keynotes, announcements, and tech content!
sasglobalforum.com | #SASGF



b_smsha
Obsidian | Level 7

I removed my city, county and name (I had merged two datasets into one so county was also one of the variables) and this worked! I was wondering though that in case I decided to increase the MAX_LEVELS, would there be a problem with the result in any modeling techniques? You mentioned that if we run regression, it may cause parameter of 999, what does this mean exactly?

MelodieRush
SAS Employee

Most modeling techniques create 1 parameter estimate for each continuous(interval) variable and n-1 parameter estimates for categorical(nominal, ordinal). 

 

For example, if income is continous then when you run a regression you will get one parameter estimate for income.  If you have a categorical variable such as gender (with the values of F, M, U), then the regression model will create 2 (n-1) parameter estimates (one for F and one for M). Most default models will use the last level as the reference level. Here's a good dicussion on reference levels in the community https://communities.sas.com/t5/Statistical-Procedures/What-Reference-Category-in-Logistic-regression... 

Catch the SAS Global Forum keynotes, announcements, and tech content!
sasglobalforum.com | #SASGF



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
  • 3 replies
  • 721 views
  • 1 like
  • 2 in conversation