Weakly supervised nonnegative matrix factorization for user-driven clustering

Jaegul Choo, Changhyun Lee, Chandan K. Reddy, Haesun Park

Research output: Contribution to journalArticle

8 Citations (Scopus)

Abstract

Clustering high-dimensional data and making sense out of its result is a challenging problem. In this paper, we present a weakly supervised nonnegative matrix factorization (NMF) and its symmetric version that take into account various prior information via regularization in clustering applications. Unlike many other existing methods, the proposed weakly supervised NMF methods provide interpretable and flexible outputs by directly incorporating various forms of prior information. Furthermore, the proposed methods maintain a comparable computational complexity to the standard NMF under an alternating nonnegativity-constrained least squares framework. By using real-world data, we conduct quantitative analyses to compare our methods against other semi-supervised clustering methods. We also present the use cases where the proposed methods lead to semantically meaningful and accurate clustering results by properly utilizing user-driven prior information.

Original languageEnglish
Pages (from-to)1598-1621
Number of pages24
JournalData Mining and Knowledge Discovery
Volume29
Issue number6
DOIs
Publication statusPublished - 2015 Nov 29
Externally publishedYes

Fingerprint

Factorization
Computational complexity

Keywords

  • Nonnegative matrix factorization
  • Regularization
  • Semi-supervised clustering
  • User-driven clustering

ASJC Scopus subject areas

  • Information Systems
  • Computer Science Applications
  • Computer Networks and Communications

Cite this

Weakly supervised nonnegative matrix factorization for user-driven clustering. / Choo, Jaegul; Lee, Changhyun; Reddy, Chandan K.; Park, Haesun.

In: Data Mining and Knowledge Discovery, Vol. 29, No. 6, 29.11.2015, p. 1598-1621.

Research output: Contribution to journalArticle

Choo, Jaegul ; Lee, Changhyun ; Reddy, Chandan K. ; Park, Haesun. / Weakly supervised nonnegative matrix factorization for user-driven clustering. In: Data Mining and Knowledge Discovery. 2015 ; Vol. 29, No. 6. pp. 1598-1621.
@article{41e3c673af97456d96c7fde21298e31b,
title = "Weakly supervised nonnegative matrix factorization for user-driven clustering",
abstract = "Clustering high-dimensional data and making sense out of its result is a challenging problem. In this paper, we present a weakly supervised nonnegative matrix factorization (NMF) and its symmetric version that take into account various prior information via regularization in clustering applications. Unlike many other existing methods, the proposed weakly supervised NMF methods provide interpretable and flexible outputs by directly incorporating various forms of prior information. Furthermore, the proposed methods maintain a comparable computational complexity to the standard NMF under an alternating nonnegativity-constrained least squares framework. By using real-world data, we conduct quantitative analyses to compare our methods against other semi-supervised clustering methods. We also present the use cases where the proposed methods lead to semantically meaningful and accurate clustering results by properly utilizing user-driven prior information.",
keywords = "Nonnegative matrix factorization, Regularization, Semi-supervised clustering, User-driven clustering",
author = "Jaegul Choo and Changhyun Lee and Reddy, {Chandan K.} and Haesun Park",
year = "2015",
month = "11",
day = "29",
doi = "10.1007/s10618-014-0384-8",
language = "English",
volume = "29",
pages = "1598--1621",
journal = "Data Mining and Knowledge Discovery",
issn = "1384-5810",
publisher = "Springer Netherlands",
number = "6",

}

TY - JOUR

T1 - Weakly supervised nonnegative matrix factorization for user-driven clustering

AU - Choo, Jaegul

AU - Lee, Changhyun

AU - Reddy, Chandan K.

AU - Park, Haesun

PY - 2015/11/29

Y1 - 2015/11/29

N2 - Clustering high-dimensional data and making sense out of its result is a challenging problem. In this paper, we present a weakly supervised nonnegative matrix factorization (NMF) and its symmetric version that take into account various prior information via regularization in clustering applications. Unlike many other existing methods, the proposed weakly supervised NMF methods provide interpretable and flexible outputs by directly incorporating various forms of prior information. Furthermore, the proposed methods maintain a comparable computational complexity to the standard NMF under an alternating nonnegativity-constrained least squares framework. By using real-world data, we conduct quantitative analyses to compare our methods against other semi-supervised clustering methods. We also present the use cases where the proposed methods lead to semantically meaningful and accurate clustering results by properly utilizing user-driven prior information.

AB - Clustering high-dimensional data and making sense out of its result is a challenging problem. In this paper, we present a weakly supervised nonnegative matrix factorization (NMF) and its symmetric version that take into account various prior information via regularization in clustering applications. Unlike many other existing methods, the proposed weakly supervised NMF methods provide interpretable and flexible outputs by directly incorporating various forms of prior information. Furthermore, the proposed methods maintain a comparable computational complexity to the standard NMF under an alternating nonnegativity-constrained least squares framework. By using real-world data, we conduct quantitative analyses to compare our methods against other semi-supervised clustering methods. We also present the use cases where the proposed methods lead to semantically meaningful and accurate clustering results by properly utilizing user-driven prior information.

KW - Nonnegative matrix factorization

KW - Regularization

KW - Semi-supervised clustering

KW - User-driven clustering

UR - http://www.scopus.com/inward/record.url?scp=84942505304&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84942505304&partnerID=8YFLogxK

U2 - 10.1007/s10618-014-0384-8

DO - 10.1007/s10618-014-0384-8

M3 - Article

AN - SCOPUS:84942505304

VL - 29

SP - 1598

EP - 1621

JO - Data Mining and Knowledge Discovery

JF - Data Mining and Knowledge Discovery

SN - 1384-5810

IS - 6

ER -