Empirical Comparison of Word Similarity Measures Based on Co-Occurrence, Context, and a Vector Space Model

다국어 초록 (Multilingual Abstract)

Word similarity is often measured to enhance system performance in the information retrieval field and other related areas. This paper reports on an experimental comparison of values for word similarity measures that were computed based on 50 intentionally selected words from a Reuters corpus. There were three targets, including (1) co-occurrence-based similarity measures (for which a co-occurrence frequency is counted as the number of documents or sentences), (2) context-based distributional similarity measures obtained from a latent Dirichlet allocation (LDA), nonnegative matrix factorization (NMF), and Word2Vec algorithm, and (3) similarity measures computed from the tf-idf weights of each word according to a vector space model (VSM). Here, a Pearson correlation coefficient for a pair of VSM-based similarity measures and co-occurrence-based similarity measures according to the number of documents was highest. Group-average agglomerative hierarchical clustering was also applied to similarity matrices computed by individual measures. An evaluation of the cluster sets according to an answer set revealed that VSM- and LDA-based similarity measures performed best.

참고문헌 (Reference)

1 Pekar, V., "Word classification based on combined measures of distributional and semantic similarity" Association for Computational Linguistics 2 : 147-150, 2003

2 Lin, H., "Topic detection from short text: A term-based consensus clustering method" IEEE 1-6, 2016

3 Peat, H. J., "The limitations of term cooccurrence data for query expansion in document retrieval systems" 42 (42): 378-383, 1991

4 Li, C. H., "Text categorization algorithms using semantic approaches, corpus-based thesaurus and WordNet" 39 (39): 765-772, 2012

5 Jo, T., "String vector based AHC as approach to word clustering" Lancaster Centre for Forecasting 133-138, 2016

6 Chen, L., "Statistical relationship determination in automatic thesaurus construction" Association for Computing Machinery 267-268, 2005

7 Dagan, I., "Similarity-based models of word cooccurrence probabilities" 34 (34): 43-69, 1999

8 Lagutina, K., "Sentiment classification of Russian texts using automatically generated thesaurus" FRUCT Oy 217-222, 2018

9 Zazo, A. F., "Reformulation of queries using similarity thesauri" 41 (41): 1163-1173, 2005

10 Xu, J., "Query expansion using local and global document analysis" Association for Computing Machinery 4-11, 1996

11 Hofmann, T., "Probabilistic latent semantic indexing" Association for Computing Machinery 50-57, 1999

12 Mohsen, G., "On the automatic construction of an Arabic thesaurus" IEEE 243-247, 2018

13 Waltz, D. L., "Massively parallel parsing : A strongly interactive model of natural language interpretation" 9 (9): 51-74, 1985

14 Ravikumar, S., "Mapping the intellectual structure of scientometrics : A co-word analysis of the journal Scientometrics(2005-2010)" 102 (102): 929-955, 2015

15 Mikolov, T., "Linguistic regularities in continuous space word representations" Association for Computational Linguistics 746-751, 2013

16 Lee, D. D., "Learning the parts of objects by non-negative matrix factorization" 401 : 788-791, 1999

17 Blei, D. M., "Latent Dirichlet allocation" 3 : 993-1022, 2003

18 Khasseh, A. A., "Intellectual structure of knowledge in iMetrics : A co-word analysis" 53 (53): 705-720, 2017

19 Deerwester, S., "Indexing by latent semantic analysis" 41 (41): 391-407, 1990

20 Gallant, S., "HNC's MatchPlus System" National Institute of Standards and Technology 107-111, 1992

21 Pennington, J., "GloVe: Global vectors for word representation" Association for Computational Linguistics 1532-1543, 2014

22 Terra, E. L., "Frequency estimates for statistical word similarity measures" Association for Computational Linguistics 1 : 165-172, 2003

23 Griffiths, T. L., "Finding scientific topics" 101 (101): 5228-5235, 2004

24 Toutanova, K., "Feature-rich part-of-speech tagging with a cyclic dependency network" Association for Computational Linguistics 173-180, 2003

25 Kishida, K., "Empirical comparison of external evaluation measures for document clustering by using synthetic data" IPSJ SIG 1-7, 2014

26 Kishida, K., "Double-pass clustering technique for multilingual document collections" 37 (37): 304-321, 2011

27 Kotlerman, L., "Directional distributional similarity for lexical inference" 16 (16): 359-389, 2010

28 Qiu, Y., "Concept based query expansion" Association for Computing Machinery 160-169, 1993

29 Mandala, R., "Combining multiple evidence from different types of thesaurus for query expansion" Association for Computing Machinery 191-197, 1999

30 Zhao, Z., "Cluster-driven model for improved word and text embedding" IOS Press 99-106, 2016

31 Poostchi, H., "Cluster labeling by word embeddings and WordNet's hypernymy" Association for Computational Linguistics 66-70, 2018

32 Xu, H., "Automatic thesaurus construction for spam filtering using revised back propagation neural network" 37 (37): 18-23, 2010

33 Liebeskind, C., "Automatic thesaurus construction for modern Hebrew" European Language Resources Association 1446-1451, 2018

34 Shunmugam, D. A., "An empirical investigation of word clustering techniques for natural language understanding" 6 (6): 2637-2646, 2016

35 Jing, Y., "An association thesaurus for information retrieval" Le Centre de Hautes Etudes Internationales d'Informatique Documentaire 1 : 146-160, 1994

36 Schutze, H., "A cooccurrence-based thesaurus and two applications to information retrieval" 33 (33): 307-318, 1997