FE: rewrote F1 evaluation using Pandas

broadinstitute / palantir-workflows

Utility workflows for the DSP hydro.gen team (formerly palantir)

BSD 3-Clause "New" or "Revised" License

19 stars 8 forks source link

Closed michaelgatzen closed 2 years ago

michaelgatzen commented 2 years ago

New (more correct) calculation of the F1 scores at each sampling position (this solves the issue of "gaps" in the plots)
Automatic detection of the stratifiers