Property type distribution in Wordnet, corpora and Wikipedia

•We present a method to compute the distribution of properties in three resources.•Other properties than taxonomic and parts can be learn from Wordnet glosses.•Corpora should be mainly used for learning quality properties.•Wikipedia should be used to extract logical statements than can be formalized...

Full description

Saved in:
Bibliographic Details
Published in:Expert systems with applications 2015-05, Vol.42 (7), p.3501-3507
Main Author: Barbu, Eduard
Format: Article
Language:English
Subjects:
Algorithms
Computation
Corpora
Expert systems
Exploration
Knowledge representation
Learning
Ontology learning
Property type distribution
Tasks
Texts
Wikipedia
Wordnet
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:•We present a method to compute the distribution of properties in three resources.•Other properties than taxonomic and parts can be learn from Wordnet glosses.•Corpora should be mainly used for learning quality properties.•Wikipedia should be used to extract logical statements than can be formalized in OWL. The ontology learning from text lacks an initial evaluation of the property types present in the resource used in the learning task. We need a way to explore the distribution of property types before the actual ontology learning process. In this paper we propose three algorithms that help ontologists in the preliminary resource exploration. The algorithms are devised for Wordnet, generic corpora and Wikipedia. Minimal assumptions about the property types that can be extracted are made. The algorithms are tested with concepts belonging to five taxonomies. The distribution of property types for the extracted concepts is computed and reported.
ISSN:0957-4174
1873-6793
DOI:10.1016/j.eswa.2014.11.070