[GRASS-user] r.learn.ml does not write the feature importance file

Hi all,

I’m rather new in using machine learning classifiers (indeed I only started to test today) and I found that setting fimp_file option does not give any output. According to the manual, it should save the feature importances in a csv file, but nothing comes out. Do I need to set permutations to a value higher than the default 10 so as to get feature importance file or how does it work?

I apologize if the question is very silly, I’m learning :wink:

Thanks in advance!

Vero

I missed the -f flag, setting that, the feature importance is estimated and written to corresponding csv file

sorry for the noise

Vero

El jue., 1 nov. 2018 a las 18:26, Veronica Andreo (<veroandreo@gmail.com>) escribió:

Hi all,

I’m rather new in using machine learning classifiers (indeed I only started to test today) and I found that setting fimp_file option does not give any output. According to the manual, it should save the feature importances in a csv file, but nothing comes out. Do I need to set permutations to a value higher than the default 10 so as to get feature importance file or how does it work?

I apologize if the question is very silly, I’m learning :wink:

Thanks in advance!

Vero