LINGUIST List 17.2666
Tue Sep 19 2006
Diss: Computational Ling: Sahlgren: 'The Word-Space Model: Using di...'
Editor for this issue: Hannah Morales
<hannahlinguistlist.org>
Directory
1. Magnus
Sahlgren,
The Word-Space Model: Using distributional analysis to represent syntagmatic and paradigmatic relations between words in high-dimensional vector spaces
Message 1: The Word-Space Model: Using distributional analysis to represent syntagmatic and paradigmatic relations between words in high-dimensional vector spaces
Date: 19-Sep-2006
From: Magnus Sahlgren <mangesics.se>
Subject: The Word-Space Model: Using distributional analysis to represent syntagmatic and paradigmatic relations between words in high-dimensional vector spaces
Institution: Stockholm University
Program: Department of Linguistics
Dissertation Status: Completed
Degree Date: 2006
Author: Magnus Sahlgren
Dissertation Title: The Word-Space Model: Using distributional analysis to represent syntagmatic and paradigmatic relations between words in high-dimensional vector spaces
Dissertation URL: http://www.sics.se/~mange/TheWordSpaceModel.pdf
Linguistic Field(s):
Computational Linguistics
Dissertation Director:
Jussi Karlgren
Dissertation Abstract:
The word-space model is a computational model of word meaning thatutilizes the distributional patterns of words collected over largetext data to represent semantic similarity between words in terms ofspatial proximity. The model has been used for over a decade, and hasdemonstrated its mettle in numerous experiments and applications. Itis now on the verge of moving from research environments to practicaldeployment in commercial systems. Although extensively used andintensively investigated, our theoretical understanding of theword-space model remains unclear. The question this dissertationattempts to answer is, 'What kind of semantic information does theword-space model acquire and represent?'
The answer is derived through an identification and discussion of thethree main theoretical cornerstones of the word-space model: thegeometric metaphor of meaning, the distributional methodology, and thestructuralist meaning theory. It is argued that the word-space modelacquires and represents two different types of relations between words- syntagmatic or paradigmatic relations - depending on how thedistributional patterns of words are used to accumulate wordspaces. The difference between syntagmatic and paradigmatic wordspaces is empirically demonstrated in a number of experiments,including comparisons with thesaurus entries, association norms, asynonym test, a list of antonym pairs, and a record of part-of-speechassignments.
|