Welcome to the InfoSci Platform
Reference Hub2
Semantic Relatedness Estimation using the Layout Information of Wikipedia Articles

Semantic Relatedness Estimation using the Layout Information of Wikipedia Articles

Patrick Chan, Yoshinori Hijikata, Toshiya Kuramochi, Shogo Nishida
Copyright: © 2013 |Volume: 7 |Issue: 2 |Pages: 19
ISSN: 1557-3958|EISSN: 1557-3966|EISBN13: 9781466632486|DOI: 10.4018/ijcini.2013040103
Cite Article Cite Article

MLA

Chan, Patrick, et al. "Semantic Relatedness Estimation using the Layout Information of Wikipedia Articles." IJCINI vol.7, no.2 2013: pp.30-48. http://doi.org/10.4018/ijcini.2013040103

APA

Chan, P., Hijikata, Y., Kuramochi, T., & Nishida, S. (2013). Semantic Relatedness Estimation using the Layout Information of Wikipedia Articles. International Journal of Cognitive Informatics and Natural Intelligence (IJCINI), 7(2), 30-48. http://doi.org/10.4018/ijcini.2013040103

Chicago

Chan, Patrick, et al. "Semantic Relatedness Estimation using the Layout Information of Wikipedia Articles," International Journal of Cognitive Informatics and Natural Intelligence (IJCINI) 7, no.2: 30-48. http://doi.org/10.4018/ijcini.2013040103

Export Reference

Mendeley
Favorite Full-Issue Download

Abstract

Computing the semantic relatedness between two words or phrases is an important problem in fields such as information retrieval and natural language processing. Explicit Semantic Analysis (ESA), a state-of-the-art approach to solve the problem uses word frequency to estimate relevance. Therefore, the relevance of words with low frequency cannot always be well estimated. To improve the relevance estimate of low-frequency words and concepts, the authors apply regression to word frequency, its location in an article, and its text style to calculate the relevance. The relevance value is subsequently used to compute semantic relatedness. Empirical evaluation shows that, for low-frequency words, the authors’ method achieves better estimate of semantic relatedness over ESA. Furthermore, when all words of the dataset are considered, the combination of the authors’ proposed method and the conventional approach outperforms the conventional approach alone.

Request Access

You do not own this content. Please login to recommend this title to your institution's librarian or purchase it from the IGI Global bookstore.