On Mon, Feb 6, 2012 at 11:34 AM, Michal Kottman
<k0mpjut0r@gmail.com> wrote:
On 6 February 2012 11:01, Francesco Abbate <francesco.bbt@gmail.com> wrote:
> 2012/2/6 Dirk Laurie <dirk.laurie@gmail.com>:
>> I've aliased `matlab` to `gsl-shell` ...
>
> Nice to hear that but pay attention, gsl-shell is not a drop in
> replacement for matlab.... :-)
Speaking of "matlab", is there a plan to add other functionality, like
machine learning algorithms (NN, SVM, ...), clustering, SVD, PCA, ...?
Does GSL support such functionality?
Or should I implement it myself when the need arises? :)
If by SVD your mean "singular value decomposition", there is already a function to get this and other solver functions will come soon.
A method for PCA ("principal component analysis" I suppose) is not in there explicitly, but can be built with the available tools and it should not be that complicated, once you have the covariance matrix.
Please let Francesco know whenever there are implementations of new features that are worth accepting into the main branch - it is always welcomed.
