Closed ivalaginja closed 5 years ago
See comments on issue #39
See this comment of the suggestion of using masked arrays when eliminating models with negative AIC.
@hrwakeford do you have example data that yields negative AIC for some of the systematic models (at least one) so that we can test this?
@hrwakeford I suppose the final output files should save all data without the mask applied to it? I will make sure to save the mask with it so that the user can reconstruct which models where actually used and which ones were masked out.
Dealing with this now over on PR #46.
Closed
pos
filters out only the systematic models with positive AIC. We need to check whether thispos
vector actually works as it should when there is at least one sys model with negative AIC, and whether this array cut propagates correctly through the rest of the code.