Multi-view cluster aggregation and splitting, with an application to multi-omic breast cancer data

Abstract : Multi-view data, which represent distinct but related groupings of variables, can be useful for identifying relevant and robust clustering structures among observations. A large number of multi-view classification algorithms have been proposed in the fields of computer science and ge-nomics; in this work, we instead focus on the task of merging or splitting an existing hard or fuzzy cluster partition based on multi-view data. This work is specifically motivated by an application involving multi-omic breast cancer data from The Cancer Genome Atlas, where multiple molecular profiles (gene expression, miRNA expression, methylation, and copy number alterations) are used to further subdivide the five currently accepted intrinsic tumor subtypes into clinically distinct subgroups of patients. In addition, we investigate the performance of the proposed multi-view splitting and aggregation algorithms, as compared to single-and concatenated-view alternatives , in a set of simulations. The multi-view splitting and aggregation algorithms developed in this work are implemented in the maskmeans R package.
Type de document :
Pré-publication, Document de travail
2018
Liste complète des métadonnées

https://hal.archives-ouvertes.fr/hal-01916941
Contributeur : Cathy Maugis-Rabusseau <>
Soumis le : jeudi 8 novembre 2018 - 21:39:55
Dernière modification le : samedi 16 mars 2019 - 01:52:43
Document(s) archivé(s) le : samedi 9 février 2019 - 14:47:59

Fichier

Multiview_Preprint.pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

  • HAL Id : hal-01916941, version 1

Citation

Antoine Godichon-Baggioni, Cathy Maugis-Rabusseau, Andrea Rau. Multi-view cluster aggregation and splitting, with an application to multi-omic breast cancer data. 2018. 〈hal-01916941〉

Partager

Métriques

Consultations de la notice

110

Téléchargements de fichiers

72