search

wkumler / squallms

Repository for the Bioconductor squallms R package
Other
3 stars 0 forks source link

Enable multinomial logistic regression to allow alternate classifications #2

Open wkumler opened 7 months ago

wkumler commented 7 months ago

While "good" and "bad" are useful classes for training, there's no real reason why we couldn't create additional classifications that are automatically categorized - e.g. "solvent peak" or "only in standards". Not sure how well the med_snr and med_cor metrics alone would be able to separate these but with additional ones (maybe sample_type?) this should be totally doable.

wkumler commented 7 months ago

I wonder if the best way to go about this is to have the user supply the regression formula? Like the default would be feat_class~med_cor+med_snr but there's no reason we should ban the user from providing their own. Would require a bunch of error handling but doesn't actually feel that difficult.