Relational Similarity Measure: An Approach Combining Wikipedia and WordNet

Abstract:

Relational similarities between two pairs of words are the degrees of their semantic relations. Vector Space Model (VSM) is used to measure the relational similarity between two pairs of words, however it needs create patterns manually and these patterns are limited. Recently, Latent Relational Analysis (LRA) is proposed and achieves state-of-art results. However, it is time-consuming and cannot express implicit semantic relations. In this study, we propose a new approach to measure relational similarities between two pairs of words by combining Wordnet3.0 and the Web-Wikipedia, thus implicit semantic relations from the very large corpus can be mined. The proposed approach mainly possesses two characters: (1) A new method is proposed in the pattern extraction step, which considers various part-of-speeches of words. (2) Wordnet3.0 is applied to calculate the semantic relatedness between a pair of words so that the implicit semantic relation of the two words can be expressed. Experimental evaluation based on the 374 SAT multiple-choice word-analogy questions, the precision of the proposed approach is 43.9%, which is lower than that of LRA suggested by Turney in 2005, but the suggested approach mainly focuses on mining the semantic relations among words.

• Periodical: Applied Mechanics and Materials (Volumes 55-57)
• Pages: 955-960
• DOI: https://doi.org/10.4028/www.scientific.net/AMM.55-57.955
• Online since: May 2011
• Authors: Yan Jiao Cao, Zhao Lu, Song Mei Cai
• Keywords: Relational Similarity, Semantic Relatedness, Wikipedia, Wordnet

References

[1] P.D. Turney and M.L. Littman, Corpus-based learning of analogies and semantic relations, Machine Learning, 60, 251–278, (2005).

DOI: 10.1007/s10994-005-0913-1

[2] P.D. Turney, Measuring semantic similarity by latent relational analysis, in Proc. of IJCAI'05, p.1136–1141, (2005).

[3] http: /en. wikipedia. org/wiki/Wikipedia.

[4] Miller, G. A. 1995. WORDNET: A Lexical Database for English. Communications of ACM (11): 39-41.

[5] Y. Li, Z. A. Bandar, and D. McLean. An Approach for Measuring Semantic Similarity between Words Using Multiple Information Sources. IEEE Trans. on Knowledge and Data Engineering, 15(4): 871–882, July/Aug. (2003).

DOI: 10.1109/tkde.2003.1209005

[6] Songmei Cai' Zhao Lu' An Improved Semantic Similarity Measure for Word Pairs'The 2010 International Conference on e-Education' e-Business' e-Management and e-Learning (IC4E 2010) Jan. 22-24' 2010' Sanya' China, 2009. IEEE.

DOI: 10.1109/ic4e.2010.20

[7] Alexander Budanitsky and Graeme Hirst. Evaluating wordnet-based measures of lexical semantic relatedness. Computational Linguistics, 32(1): 13–47, (2006).

DOI: 10.1162/coli.2006.32.1.13

[8] T. Veale, WordNet sits the sat: A knowledge-based approach to lexical analogy, in Proc. of ECAI'04, p.606–612, (2004).

[9] P.D. Turney, Similarity of semantic relations, Computational Linguistics, 32(3), 379–416, (2006).

DOI: 10.1162/coli.2006.32.3.379

[10] P.D. Turney, Expressing implicit semantic relations without supervision, in Proc. of Coling/ACL'06, p.313–320, (2006).

DOI: 10.3115/1220175.1220215

[11] D. Bollegala, Y. Matsuo, and M. Ishizuka. Www sits the sat: Measuring relational similarity on the web. In Proc. Of ECAI'08, pages 333–337, (2008).

[12] Patwardhan S, Pedersen T. Using WordNet-based context vectors to estimate the semantic relatedness of concepts. In: Proceedings of the EACL 2006 workshop, making sense of sense: Bringing computational linguistics and psycholinguistics together. Trento, Italy; 2006. p.1.