I fixed the warnings and errors that occur when computing the predictions:
one was caused because an interval over which a function is evaluated was not set right. the function would evaluate to zero over the set range. I re-centered the interval correctly.
scipy.stats's skewnorm throughs an underflow error, which doesn't really seem to be an issue: the resulting function seems fine. I am now telling numpy to ignore these errors and only raise a warning.
I fixed the warnings and errors that occur when computing the predictions: