Summary
When searching for a product on the web, a group of other similar products are proposed in advertising windows, to the user, based on previous search. These products are grouped by a similarity measure into a group with fixed size. This problem of software design for web search is a combinatorial problem called k-Cluster Problem, that consists of finding a cluster with fixed size from among a set of N items. The objective of this problem is to maximize the overall similarity among the items selected to belong to the cluster.
The challenge of this group will be to find a formulation for this problem, understand the problem complexity and obtain a solution for a set of real data.
Researcher coordinators
Graça Marques Gonçalves and Lídia Lampreia Lourenço, FCT Nova