Open chsquare opened 10 months ago
ha! good catch. I had used those for testing speed of several implementations. Probably I got confused and just let them stand here.
We can simply remove the later one, there is no improvement to be expected exactly as you said.
Hi Bene and Olaf,
I just noticed that cfg.method if-block within uf_glmfit function has two times 'matlab' and two times 'pinv' with slightly different code.
https://github.com/unfoldtoolbox/unfold/blob/caa07fdf1a881a5c31705898392df3f699461aff/src/uf_toolbox/uf_glmfit.m#L125 https://github.com/unfoldtoolbox/unfold/blob/caa07fdf1a881a5c31705898392df3f699461aff/src/uf_toolbox/uf_glmfit.m#L201
https://github.com/unfoldtoolbox/unfold/blob/caa07fdf1a881a5c31705898392df3f699461aff/src/uf_toolbox/uf_glmfit.m#L133 https://github.com/unfoldtoolbox/unfold/blob/caa07fdf1a881a5c31705898392df3f699461aff/src/uf_toolbox/uf_glmfit.m#L210
The output is probably (more or less) the same in both cases, but perhaps still something you want to fix.
Best, Christoph