BookmarkSubscribeRSS Feed
🔒 This topic is solved and locked. Need further help from the community? Please sign in and ask a new question.
hemagso
Calcite | Level 5

Hello, I am using SAS Enterprise Miner (Version 13.2) to develop gradient boosting models (Using the Arbor Procedure).

 

I am stumbling on a error that seems to be input dependent, and I am have not been able to determine what conditions can cause it. When I run the procedure on the full dataset (1570 features and 75k sample size) I get a Segmentation Violation error:

 

Segmentation Violation In Task [ ARBOR (2 ]
SptGbmClearInitialF+0x00350 - /sas92/bin/SASFoundation/9.2/sasexe/sasarbor
spt376+0x00150              - /sas92/bin/SASFoundation/9.2/sasexe/sasarbor
spt380+0x00320              - /sas92/bin/SASFoundation/9.2/sasexe/sasarbor
ArbDbu1GetNodes+0x003b0     - /sas92/bin/SASFoundation/9.2/sasexe/sasarbor
ArbTaskInputVbuf+0x001fc    - /sas92/bin/SASFoundation/9.2/sasexe/sasarbor
arb188+0x00b54              - /sas92/bin/SASFoundation/9.2/sasexe/sasarbor
arb184+0x0067c              - /sas92/bin/SASFoundation/9.2/sasexe/sasarbor
arb191+0x002f0              - /sas92/bin/SASFoundation/9.2/sasexe/sasarbor
spt308+0x00d9c              - /sas92/bin/SASFoundation/9.2/sasexe/sasarbor
zipstmt+0x0109c             - /sas92/bin/SASFoundation/9.2/sasexe/sasarbor
zipok+0x000c4               - /sas92/bin/SASFoundation/9.2/sasexe/sasarbor
yspproc+0x00830             - /sas92/bin/SASFoundation/9.2/sasexe/sasxshel
ziparse+0x000cc             - /sas92/bin/SASFoundation/9.2/sasexe/sasarbor
sasarbor+0x000a4            - /sas92/bin/SASFoundation/9.2/sasexe/sasarbor
vvtentr+0x00130             - /sas92/bin/SASFoundation/9.2/sasexe/sas
___makecontext_v2+0x00108   - /lib/sparcv9/libc.so.1

NOTE: PROCEDURE ARBOR used (Total process time):
real time 1:17.30
cpu time 21.01 seconds

NOTE: 821765 kilobytes of physical memory.
NOTE: Will use 75000 out of 75000 training cases.
NOTE: Using memory pool with 841420800 bytes.
NOTE: The SAS System stopped processing this step because of errors.
NOTE: There were 75000 observations read from the data set WORK.K_CLST_A1_75.

I had limited success in droping some features to make the code execute, but the process is guesswork based and cumbersome on my team (As we have to guess which feature may be causing the problem). Besides, I don't even know if this is a true solution.

 

I tried to reproduce the error on Dummy data, as I can't disclose the real dataset that is causing the problem, to no avail. I have a binary target and only interval inputs, and I am using the following parameters in the procedure:

 

  • MAXDEPTH: 2
  • NITERATIONS: 50
  • TRAINPROPORTION: 0.6
  • SHRINKAGE: 0.1
  • EXHAUSTIVE: 5000

Thanks in advance

1 ACCEPTED SOLUTION

Accepted Solutions
Rick_SAS
SAS Super FREQ

Anytime that you encounter a segmentation violation, you should report that issue to SAS Technical Support.

 

From the SAS log, it looks like you are running SAS 9.2, which is ancient. Is that true? You are running EM 13.2 with a SAS server that is running 9.2?

 

There have been 7 or 8 releases of SAS since 9.2. It is very possible that this problem has already been fixed in a more recent release.

View solution in original post

2 REPLIES 2
Rick_SAS
SAS Super FREQ

Anytime that you encounter a segmentation violation, you should report that issue to SAS Technical Support.

 

From the SAS log, it looks like you are running SAS 9.2, which is ancient. Is that true? You are running EM 13.2 with a SAS server that is running 9.2?

 

There have been 7 or 8 releases of SAS since 9.2. It is very possible that this problem has already been fixed in a more recent release.

hemagso
Calcite | Level 5

Thank you Rick! We tested on the new version of SAS and this resolved the issue.

SAS Innovate 2025: Save the Date

 SAS Innovate 2025 is scheduled for May 6-9 in Orlando, FL. Sign up to be first to learn about the agenda and registration!

Save the date!

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
  • 1777 views
  • 1 like
  • 2 in conversation