topic Re: Why default setting model is scored best in SAS Data Science

Why default setting model is scored best

LMAlong999 — Tue, 11 Oct 2022 23:51:02 GMT

I tried using a variable clustering node and a variable selection node to reduce the redundant variables, however, noticed that in the model comparison node, the random model scored best. why?

Re: Why default setting model is scored best

ballardw — Wed, 12 Oct 2022 01:51:43 GMT

Insufficient detail.

Define the criteria used to determine best, share the data and show the code used (or generated by nodes) to make the models.

Then someone may be able to answer.

Re: Why default setting model is scored best

CraigDeVault — Wed, 19 Oct 2022 20:51:27 GMT

As mentioned, we need more details to help answer this question. As a best practice, I want to point you to a paper that hits on several data mining topics. It is several years old, but it is a fantastic paper to help in optimizing your data mining analysis. It is called Identifying and Overcoming Common Data Mining Mistakes. It is found at the following URL:

https://support.sas.com/resources/papers/proceedings/proceedings/forum2007/073-2007.pdf