Two methods of evaluation of semantic similarity of nouns based on their modifier sets

Igor A. Bolshakov, Alexander Gelbukh

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

2 Scopus citations

Abstract

Two methods of evaluation of semantic similarity/dissimilarity of English nouns are proposed based on their modifier sets taken from Oxford Collocation Dictionary for Student of English. The first method measures similarity by the portion of modifiers commonly applicable to both nouns under evaluation. The second method measures dissimilarity by the change of the mean value of cohesion between a noun and modifiers, its own or those of the contrasted noun. Cohesion between words is measured by Stable Connection Index (SCI) based of raw Web statistics for occurrences and co-occurrences of words. It is shown that the two proposed measures are approximately in inverse monotonic dependency, while the Web evaluations confer a higher resolution.

Original languageEnglish
Title of host publicationNatural Language Processing and Information Systems - 12th International Conference on Applications of Natural Language to Information Systems, NLDB 2007, Proceedings
PublisherSpringer Verlag
Pages414-419
Number of pages6
ISBN (Print)3540733507, 9783540733508
DOIs
StatePublished - 2007
Event12th International Conference on Applications of Natural Language to Information Systems, NLDB 2007 - Paris, France
Duration: 27 Jun 200729 Jun 2007

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume4592 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference12th International Conference on Applications of Natural Language to Information Systems, NLDB 2007
Country/TerritoryFrance
CityParis
Period27/06/0729/06/07

Fingerprint

Dive into the research topics of 'Two methods of evaluation of semantic similarity of nouns based on their modifier sets'. Together they form a unique fingerprint.

Cite this