These are chat archives for numpy/numpy

28th
Jul 2017
the comment says the original version of the answer is wrong, but the author edited it
Dhruv Madeka
@dmadeka
Jul 28 2017 16:21
@ngoldbaum but that assumes the mu and Sigma you have is the mu and Sigma of the associated normal
it's not a np.log(mu) mapping
Nathan Goldbaum
@ngoldbaum
Jul 28 2017 16:25
fair enough, i haven't thought about this terribly deeply
Dhruv Madeka
@dmadeka
Jul 28 2017 16:28
haha no worries, it's kind of an irresponsible answer, it assumes the mu is the lognormal mu (and incorrectly maps) and assumes the covariance is directly the covariance of the associated normal
Nathan Goldbaum
@ngoldbaum
Jul 28 2017 19:12
well if you can get a proper multivariate lognormal into scipy I'll leave it up to you to submit a better answer :)