Hi! Sorry if this is a duplicate but I have searched through the forum and can't find an answer to this question:
I have been trying out the qiime2 sample-classifier methods as an alternative to running classifiers with different R packages.
I'm wondering why the methods implemented by this plug-in require FeatureTable[Frequency] artifacts and don't accept presence/absence or relative abundance data?
If I import a presence/absence table as a FeatureTable[Frequency] artifact the methods seem to work fine, but I am worried that I am violating assumptions somewhere in the method and can't be sure.
No you are not breaking any assumptions. This is something that has been on the to-do list to add, but just at low priority so it has not happened yet. It would be an easy fix if you would like to contribute to the q2-sample-classifier plugin source code
In the meantime, your workaround should be fine (just don't use that fake frequency table for anything else! Because for most other methods it would break assumptions!)