Mètodes i models per a l'anàlisi de clústers

Abstract

[en] In this project we study clustering methods, focusing on the two most used classes of algorithms: partitional and hierarchical. We explore definition and performance of several methods, as well as some criteria to validate results. Then, we assess computational problems arising from adapting generic clustering methods to tackle large data sets, as well as algorithms designed with the goal of improving their performance in these cases. We include R code to demonstrate practical applications of clustering and to generate graphical visualizations.