Can best subset selection method (gvselect) be used in long-format data?

[Mengmeng Li]

Hi,

I am using gvselect command followed by GEE logistic model the first time. The data is in long format with 13160 observations. The command has been run for over 20 minutes and no results output yet. Is this normal situation? I am using gvselect to select the independent variables that will be included in the GEE multiple logistic model. And the laptop I am using is Mac Air 4GB memory.

Thanks for sharing your experience. Happy Thanksgiving!

Regards,
Mengmeng

[Lukas Mer]

Hi Mengmeng,

I have no idea about the command in question but generally, it's always a good idea to run new commands on a small subsample of your data to check how they work. Especially those ML-commands are very computationally intensive - therefore its unsurprising if they take so long with your full data.

Best, Lukas