Comparison Of Local And Global Undirected Graphical Models

Zhemin Zhu; Djoerd Hiemstra; Peter M.G. Apers; Andreas Wombacher

Comparison Of Local And Global Undirected Graphical Models

Zhemin Zhu, Djoerd Hiemstra, Peter M.G. Apers, Andreas Wombacher

Databases (Former)

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

2 Citations (Scopus)

19 Downloads (Pure)

Abstract

CRFs are discriminative undirected models which are globally normalized. Global normalization preserves CRFs from the label bias problem which most local models suffer from. Recently proposed co-occurrence rate networks (CRNs) are also discriminative undirected models. In contrast to CRFs, CRNs are locally normalized. It was established that CRNs are immune to the label bias problem even they are local models. In this paper, we further compare ECRNs (using fully empirical relative frequencies, not by support vector Regression) and CRFs. The connection between Co-occurrence Rate, which is the exponential function of pointwise mutual information, and Copulas is built in continuous case. Also they are further evaluated statistically by experiments.

Original language	English
Title of host publication	ESANN 2014
Subtitle of host publication	22nd European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning : Bruges, Belgium, April 23-24-25, 2014 : proceedings
Editors	Michel Verleysen
Place of Publication	Louvain-la-Neuve, Begium
Publisher	i6doc.com
Pages	479-484
Number of pages	6
ISBN (Print)	978-2-87419-095-7
Publication status	Published - Apr 2014
Event	22nd European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning, ESANN 2014 - Bruges, Belgium Duration: 23 Apr 2014 → 25 Apr 2014

Conference

Conference	22nd European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning, ESANN 2014
Period	23/04/14 → 25/04/14
Other	23-25 April 2014

Access to Document

ESANN2014-60_3

http://www.elen.ucl.ac.be/Proceedings/esann/esannpdf/es2014-60.pdf

Cite this

Zhu, Z., Hiemstra, D., Apers, P. M. G., & Wombacher, A. (2014). Comparison Of Local And Global Undirected Graphical Models. In M. Verleysen (Ed.), ESANN 2014: 22nd European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning : Bruges, Belgium, April 23-24-25, 2014 : proceedings (pp. 479-484). i6doc.com. http://www.elen.ucl.ac.be/Proceedings/esann/esannpdf/es2014-60.pdf

Zhu, Zhemin ; Hiemstra, Djoerd ; Apers, Peter M.G. et al. / Comparison Of Local And Global Undirected Graphical Models. ESANN 2014: 22nd European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning : Bruges, Belgium, April 23-24-25, 2014 : proceedings. editor / Michel Verleysen. Louvain-la-Neuve, Begium : i6doc.com, 2014. pp. 479-484

@inproceedings{62f8f27256624a1b931caae836f17586,

title = "Comparison Of Local And Global Undirected Graphical Models",

abstract = "CRFs are discriminative undirected models which are globally normalized. Global normalization preserves CRFs from the label bias problem which most local models suffer from. Recently proposed co-occurrence rate networks (CRNs) are also discriminative undirected models. In contrast to CRFs, CRNs are locally normalized. It was established that CRNs are immune to the label bias problem even they are local models. In this paper, we further compare ECRNs (using fully empirical relative frequencies, not by support vector Regression) and CRFs. The connection between Co-occurrence Rate, which is the exponential function of pointwise mutual information, and Copulas is built in continuous case. Also they are further evaluated statistically by experiments.",

author = "Zhemin Zhu and Djoerd Hiemstra and Apers, {Peter M.G.} and Andreas Wombacher",

note = "http://www.i6doc.com/nl/livre/?GCOI=28001100432440; 22nd European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning, ESANN 2014 ; Conference date: 23-04-2014 Through 25-04-2014",

year = "2014",

month = apr,

language = "English",

isbn = "978-2-87419-095-7",

pages = "479--484",

editor = "Michel Verleysen",

booktitle = "ESANN 2014",

publisher = "i6doc.com",

address = "Belgium",

}

Zhu, Z, Hiemstra, D, Apers, PMG & Wombacher, A 2014, Comparison Of Local And Global Undirected Graphical Models. in M Verleysen (ed.), ESANN 2014: 22nd European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning : Bruges, Belgium, April 23-24-25, 2014 : proceedings. i6doc.com, Louvain-la-Neuve, Begium, pp. 479-484, 22nd European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning, ESANN 2014, 23/04/14. <http://www.elen.ucl.ac.be/Proceedings/esann/esannpdf/es2014-60.pdf>

Comparison Of Local And Global Undirected Graphical Models. / Zhu, Zhemin; Hiemstra, Djoerd; Apers, Peter M.G. et al.

ESANN 2014: 22nd European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning : Bruges, Belgium, April 23-24-25, 2014 : proceedings. ed. / Michel Verleysen. Louvain-la-Neuve, Begium : i6doc.com, 2014. p. 479-484.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

TY - GEN

T1 - Comparison Of Local And Global Undirected Graphical Models

AU - Zhu, Zhemin

AU - Hiemstra, Djoerd

AU - Apers, Peter M.G.

AU - Wombacher, Andreas

N1 - http://www.i6doc.com/nl/livre/?GCOI=28001100432440

PY - 2014/4

Y1 - 2014/4

N2 - CRFs are discriminative undirected models which are globally normalized. Global normalization preserves CRFs from the label bias problem which most local models suffer from. Recently proposed co-occurrence rate networks (CRNs) are also discriminative undirected models. In contrast to CRFs, CRNs are locally normalized. It was established that CRNs are immune to the label bias problem even they are local models. In this paper, we further compare ECRNs (using fully empirical relative frequencies, not by support vector Regression) and CRFs. The connection between Co-occurrence Rate, which is the exponential function of pointwise mutual information, and Copulas is built in continuous case. Also they are further evaluated statistically by experiments.

AB - CRFs are discriminative undirected models which are globally normalized. Global normalization preserves CRFs from the label bias problem which most local models suffer from. Recently proposed co-occurrence rate networks (CRNs) are also discriminative undirected models. In contrast to CRFs, CRNs are locally normalized. It was established that CRNs are immune to the label bias problem even they are local models. In this paper, we further compare ECRNs (using fully empirical relative frequencies, not by support vector Regression) and CRFs. The connection between Co-occurrence Rate, which is the exponential function of pointwise mutual information, and Copulas is built in continuous case. Also they are further evaluated statistically by experiments.

M3 - Conference contribution

SN - 978-2-87419-095-7

SP - 479

EP - 484

BT - ESANN 2014

A2 - Verleysen, Michel

PB - i6doc.com

CY - Louvain-la-Neuve, Begium

T2 - 22nd European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning, ESANN 2014

Y2 - 23 April 2014 through 25 April 2014

ER -

Comparison Of Local And Global Undirected Graphical Models

Abstract

Conference

Access to Document

Fingerprint

Cite this