Thursday, January 25, 2007

Clustering sets

Given sets S_1, ..., S_n, each S_i being a set of points on the plane, cluster the sets (using usual suspects of optimization criteria) with some suitable notion of distance between two pointsets. Any fun versions of this problem?

2 Comments:

Blogger Ville said...

Interesting, what do you mean by distance between point sets? If the distance is metric, at least I know EMD (Earth Mover Distance), but anythiing else?

Distances typically used agglomerative clustering do _not_ satisfy triangle inequality. I could use it to speed up my clustering algorithm. B-)

Aagh this was off-topic ... sorry.

2:07 AM  
Anonymous perfume blue said...

Ah….the sweet, smell of perfume! Today's market is flooded with hundreds and hundreds of different fragrances

ranging from floral to woodsy. Most women love the smell of perfume, wearing it even when going to the grocery

store. The problem is that perfume allergy for some women, is anything but nice.

9:33 AM  

Post a Comment

<< Home