Open F-said opened 2 years ago
I would suggest having reliability in it's own module separate from preprocess as here: https://github.com/NDCLab/pepper-pipeline/tree/dev/scripts/reliability . Other reliability metrics can then added to the same module to be re-used.
We don't actually need to write our own basic SME function since that it is the same as SEM, which there is a SciPy function for: https://docs.scipy.org/doc/scipy/reference/generated/scipy.stats.sem.html , but we can add a bootstrap based SME function there.
Background Where does
erp
andreliability
go? Preprocess/test? Should be discussed.Defined as Done Placed in a new folder or in an existing folder. Should minimize cognitive load