基于Word2Vec的WordNet词语相似度计算研究

doi:10.3778/j.issn.1002-8331.2009-0090

摘要/Abstract

摘要： 当前大部分WordNet词语相似度计算方法由于未充分考虑词语的语义信息和位置关系，导致相似度的准确率降低。为解决上述问题，提出了一种使用词向量模型Word2Vec计算WordNet词语相似度的新方法。在构建WordNet数据集时提出一种新形式，不再使用传统的文本语料库，同时提出信息位置排列方法对数据集加以处理。利用Word2Vec模型训练WordNet数据集后得到向量表示。在公开的R&G-65、M&C-30和MED38词语相似度测评集上完成了词语相似度计算任务，从多个角度进行了Pearson相关系数对比实验。结果显示该文计算的相似度值与人工判定值计算取得的Pearson相关系数指标得到了显著提升。

关键词: 词语相似度, WordNet, Word2Vec, 同义词集标号

Abstract: Currently, most WordNet word similarity calculation methods do not fully consider the semantic information and the location relationships of words, leading to the similarity accuracy reduction. To solve these problems, this paper proposes a new method to calculate the WordNet word similarity using the word vector model Word2Vec. A new form of the WordNet data set is proposed instead of using the traditional text corpus, and the information position arrangement method is used to process the data set. The vector representations are obtained by training the WordNet data set with the Word2Vec model. The word similarity calculation task is completed on the open word similarity evaluation sets like R&G-65, M&C-30 and MED38, and the Pearson correlation coefficient comparative experiment is conducted from multiple angels. Experimental results show that Pearson correlation coefficient computed by the similarity value calculated in this paper and the artificial judgement value is significantly improved.

Key words: word similarity, WordNet, Word2Vec, synset label

陈丹华, 王艳娜, 周子力, 赵晓函, 李天宇, 王凯莉. 基于Word2Vec的WordNet词语相似度计算研究[J]. 计算机工程与应用, 2022, 58(3): 222-229.

CHEN Danhua, WANG Yanna, ZHOU Zili, ZHAO Xiaohan, LI Tianyu, WANG Kaili. Research on WordNet Word Similarity Calculation Based on Word2Vec[J]. Computer Engineering and Applications, 2022, 58(3): 222-229.

参考文献

[1] ZHU X，YANG X，HUANG Y，et al.Measuring similarity and relatedness using multiple semantic relations in WordNet[J].Knowledge and Information Systems，2020，62：1539-1569.
[2] 郭小华，彭琦，邓涵，等.基于边权重的WordNet词语相似度计算[J].计算机工程与应用，2018，54（1）：172-178.
GUO X H，PENG Q，DENG H，et al.WordNet word similarity calculation based on edge weight[J].Computer Engineering and Applications，2018，54（1）：172-178.
[3] CAI Y，ZHANG Q，LU W，et al.A hybrid approach for measuring semantic similarity based on IC-weighted path distance in WordNet[J].Journal of Intelligent Information Systems，2018，51（1）：23-47.
[4] RADA R，MILI H，BICKNELL E，et al.Development and application of a metric on semantic nets[J].IEEE Transactions on System Man & Cybernetics，1989，19（1）：17-30.
[5] HIRST G，ST-ONGE D.Lexical chains as representations of context for the detection and correction of malapropisms[M]//WordNet：an electronic lexical database.Massachusetts：MIT Press，1998：305-332.
[6] WU Z，PALMER M.Verbs semantics and lexical selection[C]//Proceedings of the 32nd Annual Meeting on Association for Computational Linguistics，Stroudsburg，PA，USA，1994：133-138.
[7] LIU X，ZHOU Y，ZHENG R.Measuring semantic similarity in WordNet[C]//Proceedings of International Conference on Machine Learning and Cybernetics，Hong Kong，China，2007：3431-3435.
[8] RESNIK P.Using information content to evaluate semantic similarity in a taxonomy[C]//Proceedings of the 14th International Joint Conference on Artificial Intelligence，San Francisco，CA，USA，1995：448-453.
[9] LIN D K.An information-theoretic definition of similarity[C]//Proceedings of the 5th International Conference on Machine Learning，San Francisco，CA，USA，1998：296-304.
[10] TVERSKY A.Features of similarity[J].Readings in Cognitive Science，1988，84（4）：290-302.
[11] PETRAKIS E G M，VARELAS G，HLIAOUTAKIS A，et al.X-similarity：computing semantic similarity between concepts from different ontologies[J].Journal of Digital Information Management，2006，4（4）：233-237.
[12] RODRíGUEZ M A，EGENHOFER M J.Determining semantic similarity among entity classes from different ontologies[J].IEEE Transactions on Knowledge & Data Engineering，2013，15（2）：442-456.
[13] BANERJEE S，PEDERSEN T.Extended gloss overlaps as a measure of semantic relatedness[C]//Proceedings of the 18th International Joint Conference on Artificial Intelligence，San Francisco，CA，USA，2003：805-810.
[14] AOUICHA M B，TAIEB M A H.G2WS：gloss-based WordNet and Wiktionary semantic similarity measure[C]//2015 IEEE/ACS 12th International Conference of Computer Systems and Applications （AICCSA），Marrakech，2015：1-7.
[15] PATWARDHAN S，PEDERSEN T.Using WordNet-based context vectors to estimate the semantic relatedness of concepts[C]//Proceedings of the EACL Workshop Making Sense of Sense：Bringing Computational Linguistics and Psycholinguistics Together，2006：1-8.
[16] MIKOLOV T，CHEN K，CORRADO G，et al.Efficient estimation of word representations in vector space[J].arXiv：1301.3781，2013.
[17] 王名扬，吴欢，贾晓婷.结合word2vec与扩充情感词典的微博多元情感分类研究[J].东北师大学报（自然科学版），2019，51（1）：55-62.
WANG M Y，WU H，JIA X T.Research on multi-emotional classification of Weibo based on word2vec and extended emotional dictionary[J].Journal of Northeast Normal University（Natural Science Edition），2019，51（1）：55-62.
[18] 牛雪莹，赵恩莹.基于Word2Vec的微博文本分类研究[J].计算机系统应用，2019，28（8）：256-261.
NIU X Y，ZHAO E Y.Research on Chinese Weibo text classification based on word2vec[J].Computer Systems & Applications，2019，28（8）：256-261.
[19] JATNIKA D，BIJAKSANA M A，SURYANI A A.Word2vec model analysis for semantic similarities in English words[J].Procedia Computer Science，2019，157：160-167.
[20] FAUZI M A.Word2vec model for sentiment analysis of product reviews in indonesian language[J].International Journal of Electrical & Computer Engineering，2018，7（1）：244-252.
[21] ZHANG X，SUN S，ZHANG K.A new hybrid improved method for measuring concept semantic similarity in WordNet[J].International Arab Journal of Information Technology，2020，17（4）：433-439.
[22] MILLER G A.WordNet：a lexical database for English[J].Communications of the ACM，1995，38（11）：39-41.
[23] LEE Y Y，KE H，YEN T Y，et al.Combining and learning word embedding with WordNet for semantic relatedness and similarity measurement[J].Journal of the Association for Information Science and Technology，2020，71（6）：657-670.
[24] 詹卫东.WordNet简介[EB/OL].[2003-06].http：//ccl.pku.edu.cn/doubtfire/Course/Computa-tional%20Linguistics/contents/Intr2WordNet_zwd20030630.pdf.
ZHAN W D.Introduction to WordNet[EB/OL].[2003-06].http：//ccl.pku.edu.cn/doubtfire/Course/Computa-tional%20Linguistics/contents/Intr2WordNet_zwd20030630.pdf.
[25] ZHANG X，SUN S，ZHANG K.An information content-based approach for measuring concept semantic similarity in WordNet[J].Wireless Personal Communications，2018，103（1）：117-132.
[26] TALAFHA B，AL-SMADI M，AL-AYYOUB M，et al.Using a hierarchical softmax based on the huffman coding tree for authenticating arabic tweets[C]//2019 IEEE/ACS 16th International Conference on Computer Systems and Applications （AICCSA），2019：3-11.
[27] RUBENSTEIN H，GOODENOUGH J B.Contextual correlates of synonymy[J].Communications of the ACM，1965，8（10）：627-633.
[28] MILLER G A，CHARLES W G.Contextual correlates of semantic similarity[J].Language Cognition & Neuroscience，1991，6（1）：1-28.
[29] TAIEB M A H，AOUICHA M B，HAMADOU A B.Ontology-based approach for measuring semantic similarity[J].Engineering Applications of Artificial Intelligence，2014，36：238-261.
[30] LIU J，ZHANG Y，ZHAO Q.Adaptive vibe algorithm based on pearson correlation coefficient[C]//2019 Chinese Automation Congress（CAC），2019：4885-4889.
[31] JOULIN A，GRAVE E，BOJANOWSKI P，et al.Bag of tricks for efficient text classification[C]//Proceedings of the 15th Conference of the European Chapter of the Association for Computational Linguistics，2017：427-431.
[32] PENNINGTON J，SOCHER R，MANNING C.Glove：global vectors for word representation[C]//Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing，2014：1532-1543.