Multiplicity within clustering: challenges and unifications

Jacques-Henri Sublemontier 1
1 CA
LIFO - Laboratoire d'Informatique Fondamentale d'Orléans
Abstract : Data clustering is one of the most important unsupervised learning task and remain challenging one despite the huge amount of method proposed in the literature [J10]. The current large amount of data generated each month, days or hours have leading to the so called "Big Data" problem have made clustering one of the main tool to make further analysis applicable. We are now faced with multiple sources of information, massive and heterogeneous, coming from marketing to biology or social network analysis. The present study is concerned with the multiplicity within current clustering problem. Multiplicity can be found either in the data to analyse but also in the analysis to provide for demanding users. Thus several learning and mining paradigms have emerged since the last decade, namely multi-view clustering, consensus clustering or clustering ensemble, multiple consensus clustering or subspace and semi-supervised clustering [KZ10]. We observe here several works dedicated to these problems, then we propose a flexible framework unifying them all. The propose framework follow the collaborative clustering principle, where the objective is to find collaborative mechanisms between a set of clusterers in order to achieve different objectives related to the presented problems.
Complete list of metadatas

https://hal.inria.fr/hal-00814650
Contributor : Jacques-Henri Sublemontier <>
Submitted on : Wednesday, April 17, 2013 - 2:56:41 PM
Last modification on : Thursday, January 17, 2019 - 3:06:06 PM

Identifiers

  • HAL Id : hal-00814650, version 1

Collections

Citation

Jacques-Henri Sublemontier. Multiplicity within clustering: challenges and unifications. International Federation of Classification Societies (IFCS), Jul 2013, Tilburg, Netherlands. ⟨hal-00814650⟩

Share

Metrics

Record views

156