Related entity finding using semantic clustering based on wikipedia categories

From WikiPapers
Jump to: navigation, search

Related entity finding using semantic clustering based on wikipedia categories is a 2013 conference paper written in English by Stratogiannis G., Siolas G., Stafylopatis A. and published in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics).

[edit] Abstract

We present a system that performs Related Entity Finding, that is, Question Answering that exploits Semantic Information from the WWW and returns URIs as answers. Our system uses a search engine to gather all candidate answer entities and then a linear combination of Information Retrieval measures to choose the most relevant. For each one we look up its Wikipedia page and construct a novel vector representation based on the tokenization of the Wikipedia category names. This novel representation gives our system the ability to compute a measure of semantic relatedness between entities, even if the entities do not share any common category. We use this property to perform a semantic clustering of the candidate entities and show that the biggest cluster contains entities that are closely related semantically and can be considered as answers to the query. Performance measured on 20 topics from the 2009 TREC Related Entity Finding task shows competitive results.

[edit] References

This section requires expansion. Please, help!

Cited by

Probably, this publication is cited by others, but there are no articles available for them in WikiPapers.