Recent advances in fuzzy clustering

vista la conclusione della procedura valutativa ex art. 24 legge 240/2010, ai fini della proposta di chiamata a professore di II fascia della dott.ssa Maria Brigida Ferraro (SSD SECS-S/01, SC 13/D1), il giorno giovedì 28 marzo 2019 alle ore 12.00 in sala 34, si terrà il Seminario di Dipartimento, necessario per le successive delibere degli Organi Accademici.   In the last decades, due to the increasing complexity of data, soft clustering has received a great deal of attention. There exist different approaches that can be considered as soft. The most known is the fuzzy approach that consists in assigning objects to clusters with membership degrees ranging in the unit interval. Starting from the most known fuzzy clustering algorithm, fuzzy k-means, a large number of new algorithms have been introduced. Most of them produce a partition of objects by computing the Euclidean distance. As such, they are based on the linearity assumption and do not identify properly clusters characterized by nonlinear structures. In order to overcome this limitation, several approaches can be followed: density-based clustering, kernel-based or manifold-based, able to capture and preserve the intrinsic geometry of the data. Some new fuzzy manifold-based clustering algorithms, involving the so-called geodesic distance, are proposed and their effectiveness is shown by examples. Such algorithms are implemented in the new version of the fclust package, a useful toolbox for fuzzy clustering in R programming language. Unlike the other fuzzy clustering packages available on CRAN, fclust implements not only the fuzzy k-means but also many extensions, cluster validity indices and visualization tools. The new version allows to use fuzzy relational clustering algorithms for partitioning mixed-type data, a new version of Gustafson-Kessel algorithm to avoid singularity of the covariance matrix, fuzzy versions of some cluster comparison methods, and few minor changes as automatically selection  of number of clusters.  
Data: 
28/03/2019 - 12:00
Luogo: 
[sala 34. IV piano. Facoltà scienze statistiche. Città universitaria.]