Thumbnail - No thumbnail

Sparse k-means for mixed data via group-sparse clustering

Chavent, Marie; Lacaille, Jerome; Mourer, Alex; Olteanu, Madalina (2020), Sparse k-means for mixed data via group-sparse clustering, ESANN 2020 proceedings, European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning, i6doc.com

View/Open
Type
Communication / Conférence
External document link
https://hal.archives-ouvertes.fr/hal-03130672
Date
2020
Conference title
ESANN 2020 - 28th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning
Conference date
2020-10
Conference city
Bruges (virtuel)
Conference country
Belgium
Book title
ESANN 2020 proceedings, European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning
Publisher
i6doc.com
ISBN
978-2-87587-074-2
Metadata
Show full item record
Author(s)
Chavent, Marie cc
Institut de Mathématiques de Bordeaux [IMB]
Lacaille, Jerome
Safran Aircraft Engines
Mourer, Alex
SAMM - Statistique, Analyse et Modélisation Multidisciplinaire (SAmos-Marin Mersenne) [SAMM]
Olteanu, Madalina
CEntre de REcherches en MAthématiques de la DEcision [CEREMADE]
Abstract (EN)
The present manuscript tackles the issue of variable selection for clustering, in high dimensional data described both by numerical and categorical features. First, we build upon the sparse k-means algorithm with lasso penalty, and introduce the group-L1 penalty-already known in regression-in the unsupervised context. Second, we preprocess mixed data and transform categorical features into groups of dummy variables with appropriate scaling, on which one may then apply the group-sparse clustering procedure. The proposed method performs simultaneously clustering and feature selection, and provides meaningful partitions and meaningful features, numerical and categorical, for describing them.
Subjects / Keywords
Clustering; Kmeans algorithm; Variables selection; Sparse Models; Lasso penalty; Group lasso; Interpretability; Explainability; Weighted Kmeans