Test fail with windows and Anaconda - Low priority #1028

j-grana6 opened this Issue Aug 7, 2013 · 1 comment


None yet

2 participants

j-grana6 commented Aug 7, 2013

For some reason, testing El_regression on Windows with Anaconda fails when using NM optimization without specifying starting parameters (See Paul Hobson's test of release 0.5.0 on 8/5/2013 on the mailing list). This can be fixed by changing line 148 in statsmodels/emplike/tests/test_regression.py to

beta2ci = self.res1.conf_int_el(2, lower_bound = .59, upper_bound= 2.2, method= 'nm')

I can submit a patch but I think it might be easier for someone with privileges just to go and change the line.


Submitting a patch is the same for everyone, we all open PRs for merging, even if it's a one line change

@jseabold jseabold closed this in cbc0a45 Aug 14, 2013
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment