Sal K8U305, fredagen den 22 maj kl 13.15-14.00

Title:

Efficient algorithms for clustering

Abstract:
A clustering problem is given by a set of vertices or points S in a weighted graph, metric space or in the plane, and the objective is to find the “best” partition of S into subsets, or clusters. Typically, there is given some function measuring the quality of the partition and the number of clusters to form is given a priori. More recently, other variants of clustering without the number of clusters specified have been intensively studied. The problems of computing optimal partitions into clusters tend to be computational hard. For this reason, approximation algorithms are typically used for clustering problems. The aforementioned clustering problems have several important applications ranging from document clustering through DNA clone classification to machine learning. I will provide a short overview of clustering problems from the algorithmic perspective and my recent research within this area.

Mvh,

Mia