Stochastic subspace search for top-k multi-view clustering

Geng Li, Stephan Günnemann, Mohammed J. Zaki

Research output: Chapter in Book/Report/Conference proceedingConference contribution

2 Citations (Scopus)

Abstract

Finding multiple clustering solutions has recently gained much attention. Based on the observation that data is often multi-faceted, novel clustering methods have been introduced capable of detecting multiple, diverse clusterings. In this work-in-progress paper, we present a novel stochastic subspace search principle that tackles the requirements of multi-view clustering. The main idea is to consider each subspace as a state in a Markov chain and using Monte Carlo methods to sample the multi-view subspaces. By dynamically adapting the underlying probability density function we realize the generation of alternative clustering views. We present preliminary experimental results of our method and we describe future research directions.

Original languageEnglish
Title of host publicationMultiClust 2013 - 4th Workshop on Multiple Clusterings, Multi-View Data, and Multi-Source Knowledge-Driven Clustering, in Conj. with the 19th ACM SIGKDD Int. Conf. on KDD 2013
DOIs
Publication statusPublished - 5 Sep 2013
Externally publishedYes
EventMultiClust 2013 - 4th Workshop on Multiple Clusterings, Multi-View Data, and Multi-Source Knowledge-Driven Clustering, in Conj. with the 19th ACM SIGKDD Int. Conf. on Knowledge Discovery and Data Mining, KDD 2013 - Chicago, IL, United States
Duration: 11 Aug 201314 Aug 2013

Other

OtherMultiClust 2013 - 4th Workshop on Multiple Clusterings, Multi-View Data, and Multi-Source Knowledge-Driven Clustering, in Conj. with the 19th ACM SIGKDD Int. Conf. on Knowledge Discovery and Data Mining, KDD 2013
CountryUnited States
CityChicago, IL
Period11/8/1314/8/13

Fingerprint

Markov processes
Probability density function
Monte Carlo methods

Keywords

  • Algorithms
  • Data mining; I.5.3 [Pattern Recognition]: Clustering
  • Experimentation
  • H.2.8 [Database Management]: Database Applications

ASJC Scopus subject areas

  • Software

Cite this

Li, G., Günnemann, S., & Zaki, M. J. (2013). Stochastic subspace search for top-k multi-view clustering. In MultiClust 2013 - 4th Workshop on Multiple Clusterings, Multi-View Data, and Multi-Source Knowledge-Driven Clustering, in Conj. with the 19th ACM SIGKDD Int. Conf. on KDD 2013 https://doi.org/10.1145/2501006.2501010

Stochastic subspace search for top-k multi-view clustering. / Li, Geng; Günnemann, Stephan; Zaki, Mohammed J.

MultiClust 2013 - 4th Workshop on Multiple Clusterings, Multi-View Data, and Multi-Source Knowledge-Driven Clustering, in Conj. with the 19th ACM SIGKDD Int. Conf. on KDD 2013. 2013.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Li, G, Günnemann, S & Zaki, MJ 2013, Stochastic subspace search for top-k multi-view clustering. in MultiClust 2013 - 4th Workshop on Multiple Clusterings, Multi-View Data, and Multi-Source Knowledge-Driven Clustering, in Conj. with the 19th ACM SIGKDD Int. Conf. on KDD 2013. MultiClust 2013 - 4th Workshop on Multiple Clusterings, Multi-View Data, and Multi-Source Knowledge-Driven Clustering, in Conj. with the 19th ACM SIGKDD Int. Conf. on Knowledge Discovery and Data Mining, KDD 2013, Chicago, IL, United States, 11/8/13. https://doi.org/10.1145/2501006.2501010
Li G, Günnemann S, Zaki MJ. Stochastic subspace search for top-k multi-view clustering. In MultiClust 2013 - 4th Workshop on Multiple Clusterings, Multi-View Data, and Multi-Source Knowledge-Driven Clustering, in Conj. with the 19th ACM SIGKDD Int. Conf. on KDD 2013. 2013 https://doi.org/10.1145/2501006.2501010
Li, Geng ; Günnemann, Stephan ; Zaki, Mohammed J. / Stochastic subspace search for top-k multi-view clustering. MultiClust 2013 - 4th Workshop on Multiple Clusterings, Multi-View Data, and Multi-Source Knowledge-Driven Clustering, in Conj. with the 19th ACM SIGKDD Int. Conf. on KDD 2013. 2013.
@inproceedings{e67f3aa75c1a4ed89848106f508b837f,
title = "Stochastic subspace search for top-k multi-view clustering",
abstract = "Finding multiple clustering solutions has recently gained much attention. Based on the observation that data is often multi-faceted, novel clustering methods have been introduced capable of detecting multiple, diverse clusterings. In this work-in-progress paper, we present a novel stochastic subspace search principle that tackles the requirements of multi-view clustering. The main idea is to consider each subspace as a state in a Markov chain and using Monte Carlo methods to sample the multi-view subspaces. By dynamically adapting the underlying probability density function we realize the generation of alternative clustering views. We present preliminary experimental results of our method and we describe future research directions.",
keywords = "Algorithms, Data mining; I.5.3 [Pattern Recognition]: Clustering, Experimentation, H.2.8 [Database Management]: Database Applications",
author = "Geng Li and Stephan G{\"u}nnemann and Zaki, {Mohammed J.}",
year = "2013",
month = "9",
day = "5",
doi = "10.1145/2501006.2501010",
language = "English",
isbn = "9781450323345",
booktitle = "MultiClust 2013 - 4th Workshop on Multiple Clusterings, Multi-View Data, and Multi-Source Knowledge-Driven Clustering, in Conj. with the 19th ACM SIGKDD Int. Conf. on KDD 2013",

}

TY - GEN

T1 - Stochastic subspace search for top-k multi-view clustering

AU - Li, Geng

AU - Günnemann, Stephan

AU - Zaki, Mohammed J.

PY - 2013/9/5

Y1 - 2013/9/5

N2 - Finding multiple clustering solutions has recently gained much attention. Based on the observation that data is often multi-faceted, novel clustering methods have been introduced capable of detecting multiple, diverse clusterings. In this work-in-progress paper, we present a novel stochastic subspace search principle that tackles the requirements of multi-view clustering. The main idea is to consider each subspace as a state in a Markov chain and using Monte Carlo methods to sample the multi-view subspaces. By dynamically adapting the underlying probability density function we realize the generation of alternative clustering views. We present preliminary experimental results of our method and we describe future research directions.

AB - Finding multiple clustering solutions has recently gained much attention. Based on the observation that data is often multi-faceted, novel clustering methods have been introduced capable of detecting multiple, diverse clusterings. In this work-in-progress paper, we present a novel stochastic subspace search principle that tackles the requirements of multi-view clustering. The main idea is to consider each subspace as a state in a Markov chain and using Monte Carlo methods to sample the multi-view subspaces. By dynamically adapting the underlying probability density function we realize the generation of alternative clustering views. We present preliminary experimental results of our method and we describe future research directions.

KW - Algorithms

KW - Data mining; I.5.3 [Pattern Recognition]: Clustering

KW - Experimentation

KW - H.2.8 [Database Management]: Database Applications

UR - http://www.scopus.com/inward/record.url?scp=84883288864&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84883288864&partnerID=8YFLogxK

U2 - 10.1145/2501006.2501010

DO - 10.1145/2501006.2501010

M3 - Conference contribution

AN - SCOPUS:84883288864

SN - 9781450323345

BT - MultiClust 2013 - 4th Workshop on Multiple Clusterings, Multi-View Data, and Multi-Source Knowledge-Driven Clustering, in Conj. with the 19th ACM SIGKDD Int. Conf. on KDD 2013

ER -