東華大學圖書館 |

Language: English

Help

回圖書館首頁

手機版館藏查詢

Back

Switch To: Labeled | MARC Mode | ISBD

Improving hierarchical multiclass pe...

The University of Texas at Dallas.

Linked to FindBook

Google Book

Amazon

博客來

Improving hierarchical multiclass perceptrons for entity detection using Google Sets and pronoun disambiguation.

Record Type:	Electronic resources : Monograph/item
Title/Author:	Improving hierarchical multiclass perceptrons for entity detection using Google Sets and pronoun disambiguation./
Author:	Quinn, Michael.
Description:	61 p.
Notes:	Adviser: Vincent Ng.
Contained By:	Masters Abstracts International46-05.
Subject:	Computer Science. -
Online resource:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=1452726
ISBN:	9780549526667

Improving hierarchical multiclass perceptrons for entity detection using Google Sets and pronoun disambiguation.
Quinn, Michael.

Improving hierarchical multiclass perceptrons for entity detection using Google Sets and pronoun disambiguation. - 61 p.

Adviser: Vincent Ng.

Thesis (M.S.C.S.)--The University of Texas at Dallas, 2008.

Entity detection is the process of identifying objects in a body of text and correctly classifying each object into one of a set of predefined types. I propose two extensions to improve the performance of hierarchical multiclass perceptrons when used to solve entity detection. The first extension is to use Google Sets to automatically generate training instances. This will address the problems of too few training instances and high unseen word rates. The second extension is to disambiguate pronouns by identifying the influences on the pronoun classification, and using those influences as features. This will address the problem that pronouns carry less information than nominals and proper nouns. A noticeable improvement in F-measure was seen using the combination of these two approaches.

ISBN: 9780549526667Subjects--Topical Terms:

626642
Computer Science.

Improving hierarchical multiclass perceptrons for entity detection using Google Sets and pronoun disambiguation.
LDR:01675nmm 2200265 a 45 001 886344
005 20101007
008 101007s2008 ||||||||||||||||| ||eng d
020 $a 9780549526667
035 $a (UMI)AAI1452726
035 $a AAI1452726
040 $a UMI $c UMI
100 1 $a Quinn, Michael. $3 689458
245 1 0 $a Improving hierarchical multiclass perceptrons for entity detection using Google Sets and pronoun disambiguation.
300 $a 61 p.
500 $a Adviser: Vincent Ng.
500 $a Source: Masters Abstracts International, Volume: 46-05, page: 2732.
502 $a Thesis (M.S.C.S.)--The University of Texas at Dallas, 2008.
520 $a Entity detection is the process of identifying objects in a body of text and correctly classifying each object into one of a set of predefined types. I propose two extensions to improve the performance of hierarchical multiclass perceptrons when used to solve entity detection. The first extension is to use Google Sets to automatically generate training instances. This will address the problems of too few training instances and high unseen word rates. The second extension is to disambiguate pronouns by identifying the influences on the pronoun classification, and using those influences as features. This will address the problem that pronouns carry less information than nominals and proper nouns. A noticeable improvement in F-measure was seen using the combination of these two approaches.
590 $a School code: 0382.
650 4 $a Computer Science. $3 626642
690 $a 0984
710 2 $a The University of Texas at Dallas. $3 1018411
773 0 $t Masters Abstracts International $g 46-05.
790 $a 0382
790 1 0 $a Ng, Vincent, $e advisor
791 $a M.S.C.S.
792 $a 2008
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=1452726