Section outline
-
LECTURES N.25 and 26 and 27
clustering (HA, 2.6 hierarchical vs partitioning methods, see also Altman2017).
distances, metric spaces
density based (topological) vs coupling (interaction) based methods
k-means, Dbscan, superparamagnetic (Domany2003)
information based methods (Bialek2004, Slonim2005, Luksza2010)
comparing clusterings: the mutual information approach (Meila2007, Vinh2010)
Partitioning clusterings as random fragmentation processes: breaking of self-averaging
(see: Andrea De Martino, The Geometrically Broken Object, 1998 https://arxiv.org/abs/cond-mat/9805204)
STUDY MATERIALS IN CB_23_24_PACK_10-
Students mustMark as done
Find in the PACK the papers referred to in the discussion and pdf files of the lectures on clustering that were used in the CB course of 2021.
-