[LivingNightmare]

Can anyone recommend a good machine learning library in C++? The only real requirement I need it to have, is that it must support unsupervised clustering on n-dimensional data, using a user-defined comparison metric. If anyone has every used a good machine learning lib in C++, I would love to hear recommendations.

Let me know,

L.N

[Skydiver]

Any particular reason why it has to be in C++? I've heard that LISP is the language of choice for most AI research.

[modi123_1]

Not really skydiver.. I've done plenty in c++, java, and event some C#.

[Skydiver]

I guess my age is showing.

Most of the C++ AI I've seen has either been purpose built for SPAM detection or game logic.

[LivingNightmare]

It's because the existing code base I have for the project is in C++, and that's the language I feel most comfortable with. Although I've already done a good chunk of coding in C++, what I wrote is mostly for pre-processing things, so I do have a little flexibility.

Basically, the goal of this project is to try and cluster webpages based on a topic-scheme,
and then partition these centroids on different nodes in order to improve search engine queries. Instead of having the entire cluster of nodes light up in order to perform a search, if we can determine the possible topics of the user's query, then only a few nodes would be required to light up - It's for a URA project I'm doing with one of the profs at my University. Since my original post, I have also come across this in Java: http://mallet.cs.umass.edu/topics.php, which seems pretty solid. The only problem is, this looks more of a supervised method, because I need to specify the number of topics (which I do not know ahead of time), along with a few other parameters.

So I guess the actual clustering can be in any language, especially if it will work well for what I need it for (which is the unsupervised clustering/topic-seperation part).

This post has been edited by LivingNightmare: 09 August 2012 - 02:43 PM