Evaluation of feature combination for effective structural disambiguation

So Young Park; Yong Jae Kwak; Joon Ho Lim; Hae Chang Rim

Evaluation of feature combination for effective structural disambiguation

So Young Park, Yong Jae Kwak, Joon Ho Lim, Hae Chang Rim

* Honorary professors

Research output: Contribution to journal › Article › peer-review

Abstract

In this paper, we present the useful features of a syntactic constituent for a probabilistic parsing model and analyze the combination of the features in order to disambiguate parse trees effectively. Unlike most of previous works focusing on the features of a single head, the features of a functional head, the features of a content head, and the features of size are utilized in this study. Experimental results show that the combination of different features such as the functional head feature and the size feature is prefered to the combination of similar features such as the functional head feature and the content head feature. Besides, it is remarkable that the function feature is more useful than the combination of the content feature and the size feature.

Original language	English
Pages (from-to)	97-101
Number of pages	5
Journal	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	2945
Publication status	Published - 2004

ASJC Scopus subject areas

Theoretical Computer Science
Computer Science(all)

Cite this

Evaluation of feature combination for effective structural disambiguation. / Park, So Young; Kwak, Yong Jae; Lim, Joon Ho; Rim, Hae Chang.

In: Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), Vol. 2945, 2004, p. 97-101.

Research output: Contribution to journal › Article › peer-review

@article{cfc7265e1c6d465ab90ad0106137fa15,

title = "Evaluation of feature combination for effective structural disambiguation",

abstract = "In this paper, we present the useful features of a syntactic constituent for a probabilistic parsing model and analyze the combination of the features in order to disambiguate parse trees effectively. Unlike most of previous works focusing on the features of a single head, the features of a functional head, the features of a content head, and the features of size are utilized in this study. Experimental results show that the combination of different features such as the functional head feature and the size feature is prefered to the combination of similar features such as the functional head feature and the content head feature. Besides, it is remarkable that the function feature is more useful than the combination of the content feature and the size feature.",

author = "Park, {So Young} and Kwak, {Yong Jae} and Lim, {Joon Ho} and Rim, {Hae Chang}",

year = "2004",

language = "English",

volume = "2945",

pages = "97--101",

journal = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

issn = "0302-9743",

publisher = "Springer Verlag",

}

TY - JOUR

T1 - Evaluation of feature combination for effective structural disambiguation

AU - Park, So Young

AU - Kwak, Yong Jae

AU - Lim, Joon Ho

AU - Rim, Hae Chang

PY - 2004

Y1 - 2004

N2 - In this paper, we present the useful features of a syntactic constituent for a probabilistic parsing model and analyze the combination of the features in order to disambiguate parse trees effectively. Unlike most of previous works focusing on the features of a single head, the features of a functional head, the features of a content head, and the features of size are utilized in this study. Experimental results show that the combination of different features such as the functional head feature and the size feature is prefered to the combination of similar features such as the functional head feature and the content head feature. Besides, it is remarkable that the function feature is more useful than the combination of the content feature and the size feature.

AB - In this paper, we present the useful features of a syntactic constituent for a probabilistic parsing model and analyze the combination of the features in order to disambiguate parse trees effectively. Unlike most of previous works focusing on the features of a single head, the features of a functional head, the features of a content head, and the features of size are utilized in this study. Experimental results show that the combination of different features such as the functional head feature and the size feature is prefered to the combination of similar features such as the functional head feature and the content head feature. Besides, it is remarkable that the function feature is more useful than the combination of the content feature and the size feature.

UR - http://www.scopus.com/inward/record.url?scp=35048823897&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=35048823897&partnerID=8YFLogxK

M3 - Article

AN - SCOPUS:35048823897

VL - 2945

SP - 97

EP - 101

JO - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

JF - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SN - 0302-9743

ER -

Evaluation of feature combination for effective structural disambiguation

Abstract

ASJC Scopus subject areas

Other files and links

Fingerprint

Cite this