Unsupervised cross-domain learning by interaction information co-clustering

Shin Ando, Einoshin Suzuki

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

4 Citations (Scopus)

Abstract

In real-world data mining applications, one often has access to multiple datasets that are relevant to the task at hand. However, learning from such datasets can be difficult as they are often drawn from different domains, i.e., not identically distributed or differ in class or feature sets. In this paper, we consider the problem of learning the class structures of related domains in an unsupervised manner. Its setting generalizes that of information filtering and novelty detection applications which addresses both known and unknown classes. We propose a co-clustering framework for estimating and adapting the class structures of two related domains, enabling the analyses of shared and unique classes. We define an objective function using interaction information to take account of the divergence between the corresponding clusters of respective domains. We present an iterative algorithm which alternates object and feature clustering and converges to a local minimum of the objective function. We present empirical results using text benchmarks, comparing the proposed algorithm and combinations of conventional approaches in problems of partitioning documents and detecting unknown topics.

Original languageEnglish
Title of host publicationProceedings - 8th IEEE International Conference on Data Mining, ICDM 2008
Pages13-22
Number of pages10
DOIs
Publication statusPublished - Dec 1 2008
Event8th IEEE International Conference on Data Mining, ICDM 2008 - Pisa, Italy
Duration: Dec 15 2008Dec 19 2008

Publication series

NameProceedings - IEEE International Conference on Data Mining, ICDM
ISSN (Print)1550-4786

Other

Other8th IEEE International Conference on Data Mining, ICDM 2008
CountryItaly
CityPisa
Period12/15/0812/19/08

Fingerprint

Information filtering
Data mining

All Science Journal Classification (ASJC) codes

  • Engineering(all)

Cite this

Ando, S., & Suzuki, E. (2008). Unsupervised cross-domain learning by interaction information co-clustering. In Proceedings - 8th IEEE International Conference on Data Mining, ICDM 2008 (pp. 13-22). [4781096] (Proceedings - IEEE International Conference on Data Mining, ICDM). https://doi.org/10.1109/ICDM.2008.92

Unsupervised cross-domain learning by interaction information co-clustering. / Ando, Shin; Suzuki, Einoshin.

Proceedings - 8th IEEE International Conference on Data Mining, ICDM 2008. 2008. p. 13-22 4781096 (Proceedings - IEEE International Conference on Data Mining, ICDM).

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

Ando, S & Suzuki, E 2008, Unsupervised cross-domain learning by interaction information co-clustering. in Proceedings - 8th IEEE International Conference on Data Mining, ICDM 2008., 4781096, Proceedings - IEEE International Conference on Data Mining, ICDM, pp. 13-22, 8th IEEE International Conference on Data Mining, ICDM 2008, Pisa, Italy, 12/15/08. https://doi.org/10.1109/ICDM.2008.92
Ando S, Suzuki E. Unsupervised cross-domain learning by interaction information co-clustering. In Proceedings - 8th IEEE International Conference on Data Mining, ICDM 2008. 2008. p. 13-22. 4781096. (Proceedings - IEEE International Conference on Data Mining, ICDM). https://doi.org/10.1109/ICDM.2008.92
Ando, Shin ; Suzuki, Einoshin. / Unsupervised cross-domain learning by interaction information co-clustering. Proceedings - 8th IEEE International Conference on Data Mining, ICDM 2008. 2008. pp. 13-22 (Proceedings - IEEE International Conference on Data Mining, ICDM).
@inproceedings{9b30f6fbc6374d8db09d62663809b85d,
title = "Unsupervised cross-domain learning by interaction information co-clustering",
abstract = "In real-world data mining applications, one often has access to multiple datasets that are relevant to the task at hand. However, learning from such datasets can be difficult as they are often drawn from different domains, i.e., not identically distributed or differ in class or feature sets. In this paper, we consider the problem of learning the class structures of related domains in an unsupervised manner. Its setting generalizes that of information filtering and novelty detection applications which addresses both known and unknown classes. We propose a co-clustering framework for estimating and adapting the class structures of two related domains, enabling the analyses of shared and unique classes. We define an objective function using interaction information to take account of the divergence between the corresponding clusters of respective domains. We present an iterative algorithm which alternates object and feature clustering and converges to a local minimum of the objective function. We present empirical results using text benchmarks, comparing the proposed algorithm and combinations of conventional approaches in problems of partitioning documents and detecting unknown topics.",
author = "Shin Ando and Einoshin Suzuki",
year = "2008",
month = "12",
day = "1",
doi = "10.1109/ICDM.2008.92",
language = "English",
isbn = "9780769535029",
series = "Proceedings - IEEE International Conference on Data Mining, ICDM",
pages = "13--22",
booktitle = "Proceedings - 8th IEEE International Conference on Data Mining, ICDM 2008",

}

TY - GEN

T1 - Unsupervised cross-domain learning by interaction information co-clustering

AU - Ando, Shin

AU - Suzuki, Einoshin

PY - 2008/12/1

Y1 - 2008/12/1

N2 - In real-world data mining applications, one often has access to multiple datasets that are relevant to the task at hand. However, learning from such datasets can be difficult as they are often drawn from different domains, i.e., not identically distributed or differ in class or feature sets. In this paper, we consider the problem of learning the class structures of related domains in an unsupervised manner. Its setting generalizes that of information filtering and novelty detection applications which addresses both known and unknown classes. We propose a co-clustering framework for estimating and adapting the class structures of two related domains, enabling the analyses of shared and unique classes. We define an objective function using interaction information to take account of the divergence between the corresponding clusters of respective domains. We present an iterative algorithm which alternates object and feature clustering and converges to a local minimum of the objective function. We present empirical results using text benchmarks, comparing the proposed algorithm and combinations of conventional approaches in problems of partitioning documents and detecting unknown topics.

AB - In real-world data mining applications, one often has access to multiple datasets that are relevant to the task at hand. However, learning from such datasets can be difficult as they are often drawn from different domains, i.e., not identically distributed or differ in class or feature sets. In this paper, we consider the problem of learning the class structures of related domains in an unsupervised manner. Its setting generalizes that of information filtering and novelty detection applications which addresses both known and unknown classes. We propose a co-clustering framework for estimating and adapting the class structures of two related domains, enabling the analyses of shared and unique classes. We define an objective function using interaction information to take account of the divergence between the corresponding clusters of respective domains. We present an iterative algorithm which alternates object and feature clustering and converges to a local minimum of the objective function. We present empirical results using text benchmarks, comparing the proposed algorithm and combinations of conventional approaches in problems of partitioning documents and detecting unknown topics.

UR - http://www.scopus.com/inward/record.url?scp=67049100148&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=67049100148&partnerID=8YFLogxK

U2 - 10.1109/ICDM.2008.92

DO - 10.1109/ICDM.2008.92

M3 - Conference contribution

AN - SCOPUS:67049100148

SN - 9780769535029

T3 - Proceedings - IEEE International Conference on Data Mining, ICDM

SP - 13

EP - 22

BT - Proceedings - 8th IEEE International Conference on Data Mining, ICDM 2008

ER -