BookmarkSubscribeRSS Feed
🔒 This topic is solved and locked. Need further help from the community? Please sign in and ask a new question.
elomqg
Calcite | Level 5
When I run the gradient boosting model in SAS enterprise miner, it only does one iteration even though I specified 100 iterations. None of the variables turn out to be important. The root mean square error for both validation and training was a vertical line at the exact same point(overlapping each other). I get this weird result when I use 40% or more observations as training. When I use a smaller percentage of observation as my training set, I don't have this issue. I have tried other modelling techniques and get regular results irrespective of varying training/validation sample sizes. My dataset has about 400,000 observations. Any thoughts? Suggestions?
 
1 ACCEPTED SOLUTION

Accepted Solutions
PadraicGNeville
SAS Employee

A similar situation was recently resolved by specifying LEAFFRACTION to small number, such as 0.001.

View solution in original post

2 REPLIES 2
PadraicGNeville
SAS Employee

Yes this is weird.  I guess no split is created in the initial tree.   I cannot guess why reducing the number of training observations fixes it.  If you are willing and able to provide the data to SAS Tech Support, I will figure out why this is happening.

-Padraic

PadraicGNeville
SAS Employee

A similar situation was recently resolved by specifying LEAFFRACTION to small number, such as 0.001.

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
  • 2 replies
  • 1502 views
  • 0 likes
  • 2 in conversation