In recent years we are witnessing to an increased attention towards methods for clustering matrix-valued data. In this framework, matrix Gaussian mixture models constitute a natural extension of the model-based clustering strategies. Regrettably, the overparametrization issues, already affecting the vector-valued framework in high-dimensional scenarios, are even more troublesome for matrix mixtures. In this work we introduce a sparse model-based clustering procedure conceived for the matrix-variate context. We introduce a penalized estimation scheme which, by shrinking some of the parameters towards zero, produces parsimonious solutions when the dimensions increase. Moreover it allows cluster-wise sparsity, possibly easing the interpretation and providing richer insights on the analyzed dataset.
Model-based clustering with sparse matrix mixture models / A. Cappozzo, A. Casa, M. Fop (PROCEEDINGS E REPORT). - In: CLADAG 2021 / [a cura di] G. Porzio, C. Rampichini, C. Bocci. - [s.l] : Firenze University Press, 2021. - ISBN 978-88-5518-340-6. - pp. 280-283 (( Intervento presentato al 13. convegno Scientific Meeting Classification and Data Analysis Group tenutosi a Firenze nel 2021.
Model-based clustering with sparse matrix mixture models
A. Cappozzo;
2021
Abstract
In recent years we are witnessing to an increased attention towards methods for clustering matrix-valued data. In this framework, matrix Gaussian mixture models constitute a natural extension of the model-based clustering strategies. Regrettably, the overparametrization issues, already affecting the vector-valued framework in high-dimensional scenarios, are even more troublesome for matrix mixtures. In this work we introduce a sparse model-based clustering procedure conceived for the matrix-variate context. We introduce a penalized estimation scheme which, by shrinking some of the parameters towards zero, produces parsimonious solutions when the dimensions increase. Moreover it allows cluster-wise sparsity, possibly easing the interpretation and providing richer insights on the analyzed dataset.File | Dimensione | Formato | |
---|---|---|---|
published_cladag_2021_sparsemixmat.pdf
accesso aperto
Tipologia:
Publisher's version/PDF
Dimensione
471.31 kB
Formato
Adobe PDF
|
471.31 kB | Adobe PDF | Visualizza/Apri |
Pubblicazioni consigliate
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.