[mlpack-git] [mlpack] segmentation fault with LogisticRegression API (#428)

[John Lees]

Hi again Ryan,

Thanks for the tips about using 64 bit words, I certainly got myself into something bad there...
I followed the linear algebra through properly, and it would seem I needed to take the transpose of the design matrix x. Having done this before passing to the mlpack function, all works as expected!

I'm not sure whether this is my oversight, or that the mlpack function does not follow convention?
(to recap, I was supplying predictors with 3069 rows and 1 column, and equally dimensioned responses. If I take the transpose of the responses so they're 1x3069 instead all is well)
My reading of https://en.wikipedia.org/wiki/Design_matrix suggests the latter, but I might well have got that wrong.

Either way, perhaps a test + error on the initial dimensions would be preferable to a segfault?

Thanks again for all your help, and for a great bit of software!

---
Reply to this email directly or view it on GitHub:
https://github.com/mlpack/mlpack/issues/428#issuecomment-88130711
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.cc.gatech.edu/pipermail/mlpack-git/attachments/20150331/b3ef55c6/attachment.html>