[mlpack-git] [mlpack] Implemented multi-dimensional scaling (#430)

[Ryan Curtin]

Sorry for the slow response here.  I hope to have a chance to look soon, but I can't promise anything with paper deadlines coming up.  I think maybe we can expand on the tests a little bit; could you consider corner cases?  For example, what happens if a user passes an invalid number of dimensions or invalid data to the reduce() function?

Another useful test to add might be to take a dataset and pad it with many dimensions which are all zero.  In this case, you should be able to reduce it to the original number of dimensions while preserving the exact dissimilarity matrix.

If you could please also take a look through the style guidelines (which are mostly complete), and make those changes, it would be helpful. :)  https://github.com/mlpack/mlpack/wiki/StyleGuidelines

---
Reply to this email directly or view it on GitHub:
https://github.com/mlpack/mlpack/pull/430#issuecomment-99961850
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.cc.gatech.edu/pipermail/mlpack-git/attachments/20150507/65294a3d/attachment.html>