SCIENCE CHINA Information Sciences, Volume 60, Issue 11: 110101(2017) https://doi.org/10.1007/s11432-016-9229-y

## Encoding syntactic representations with a neural network for sentiment collocation extraction

Bing QIN2,*,
• AcceptedMay 8, 2017
• PublishedOct 16, 2017
Share
Rating

### Abstract

Sentiment collocation refers to the collocation of a target word and a polarity word.Sentiment collocation extraction aims to extract the targets and their modifying polarity words by analyzing the relationships between them.This can be regarded as a basic sentiment analysis task and is relevant in many practical applications.Previous studies relied mainly on the syntactic path, which is used to connect the target word and the polarity word.To deeply exploit the semantic information of the syntactic path,we propose two types of syntactic representation, namely, relation embedding and subtree embedding,to capture the latent semantic features.Relation embedding is used to represent the latent semantics between targets and their corresponding polarity words,and subtree embedding is used to explore the rich syntactic information for each word on the path.To combine the two types of syntactic representations,a neural network is constructed.We use a recursive neural network (RNN) to model the subtree embeddings, andcolorblack

### References

[1] Pang B, Lee L. Opinion Mining and Sentiment Analysis. FNT Inf Retrieval, 2008, 2: 1-135 CrossRef Google Scholar

[2] Liu B. Sentiment analysis and opinion mining. Synth Lect Human Language Tech, 2012, 5: 1--167. Google Scholar

[3] Abbasi A, Chen H, Salem A. Sentiment analysis in multiple languages: feature selection for opinion classification in web forums. ACM Trans Inf Syst, 2008, 26: 1--34. Google Scholar

[4] Duric A, Song F. Feature selection for sentiment analysis based on content and syntax models. Decision Support Syst, 2012, 53: 704-711 CrossRef Google Scholar

[5] Bloom K, Garg N, Argamon S. Extracting appraisal expressions. In: Proceedings of Human Language Technologies: the Annual Conference of the North American Chapter of the Association for Computational Linguistics, Rochester, 2007. 308--315. Google Scholar

[6] Qiu G, Liu B, Bu J J, et al. Opinion word expansion and target extraction through double propagation. Comput Linguist, 2011, 37: 9--27. Google Scholar

[7] Xu L H, Liu K, Lai S, et al. Mining opinion words and opinion targets in a two-stage framework. In: Proceedings of the 51st Annual Meeting of the Association for Computational Linguistics, Sofia, 2013. 1764--1773. Google Scholar

[8] Zhao Y Y, Che W X, Guo H L, et al. Sentence compression for target-polarity word collocation extraction. In: Proceedings of the 25th International Conference on Computational Linguistics: Technical Papers, Dublin, 2014. 1360--1369. Google Scholar

[9] Zhao Y Y, Li S Q, Qin B, et al. Encoding dependency representation with convolutional neural network for target-polarityword collocation extraction. In: Social Meida Processing. Singapore: Springer, 2016. Google Scholar

[10] Chen Y B, Xu L, Liu K, et al. Event extraction via dynamic multi-pooling convolutional neural networks. In: Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing, Beijing, 2015. 167--176. Google Scholar

[11] Liu Y, Wei F R, Li S J, et al. A dependency-based neural network for relation classification. In: Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing, Beijing, 2015. 285--290. Google Scholar

[12] Meng F D, Lu Z D, Wang M X, et al. Encoding source language with convolutional neural network for machine translation. In: Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing, Beijing, 2015. 20--30. Google Scholar

[13] Nguyen T H, Grishman R. Event detection and domain adaptation with convolutional neural networks. In: Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing, Beijing, 2015. 365--371. Google Scholar

[14] Vu N T, Adel H, Gupta P, et al. Combining recurrent and convolutional neural networks for relation classification. In: Proceedings of the 2016 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, San Diego, 2016. 534--539. Google Scholar

[15] Lee J Y, Dernoncourt F. Sequential short-text classification with recurrent and convolutional neural networks. In: Proceedings of the 2016 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, San Diego, 2016. 515--520. Google Scholar

[16] Zhao J, Xu H B, Huang X J, et al. Overview of chinese pinion analysis evaluation 2008. Proceedings of the 1st Chinese Opinion Analysis Evaluation (COAE), 2008. Google Scholar

[17] Che W X, Li Z H, Liu T. LTP: a chinese language technology platform. In: Proceedings of the 23rd International Conference on Computational Linguistics: Demonstrations, Beijing, 2010. 13--16. Google Scholar

[18] Joachims T. Learning to Classify Text Using Support Vector Machines -- Methods, Theory, and Algorithms. Norwell: Kluwer Academic Publishers, 2002. Google Scholar

[19] Gui L, Zhou Y, Xu R F, et al. Learning representations from heterogeneous network for sentiment classification of product reviews. Knowledge-Based Syst, 2017, 124: 34--45. Google Scholar

[20] Chen T, Xu R, He Y. Learning User and Product Distributed Representations Using a Sequence Model for Sentiment Analysis. IEEE Comput Intell Mag, 2016, 11: 34-44 CrossRef Google Scholar

[21] Hu M Q, Liu B. Mining and summarizing customer reviews. In: Proceedings of the 10th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, Seattle, 2004. 168--177. Google Scholar

[22] dos Santos C, Xiang B, Zhou B. Classifying relations by ranking with convolutional neural networks. In: Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing, Beijing, 2015. 626--634. Google Scholar

[23] Kim Y. Convolutional neural networks for sentence classification. In: Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP), Doha, 2014. 1746--1751. Google Scholar

[24] Ma M B, Huang L, Xiang B, et al. Dependency-based convolutional neural networks for sentence embedding. In: Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing, Beijing, 2015. 174--179. Google Scholar

• Table 1   Statistics for the Chinese datasets of six domains
 Domain # Reviews # Sentences # Collocations Camera 138 1249 1335 Notebook 56 623 674 Phone 123 1350 1479 Book 349 1270 424 Hotel 440 1808 1120 Restaurant 386 1398 689 All 1492 7698 5721
• Table 2   Results of sentiment collocation extraction using different methods $^{\rm~a)}$
 Domain Method $P$ (%) $R$ (%) $F$-score (%) Domain Method $P$ (%) $R$ (%) $F$-score (%) CNN_PathWord 72.1 48.0 57.8 CNN_PathWord 63.0 47.0 54.1 CNN_PW_Relation 74.3 77.0 75.7 CNN_PW_Relation 68.2 76.0 71.9 Phone CNN_PW_R_Flat 74.7 80.0 77.3 Camera CNN_PW_R_Flat 69.1 77.0 73.0 CNN_PW_R_Subtree 76.3 81.0 78.5 CNN_PW_R_Subtree 71.8 75.0 73.6 Feature-Based 60.6 83.9 70.3 Feature-Based 52.0 75.2 61.5 Path-Rule-Based (*) 77.3 60.9 68.1 Path-Rule-Based (*) 74.7 58.4 65.6 CNN_PathWord 62.1 48.0 54.4 CNN_PathWord 49.3 100.0 66.0 CNN_PW_Relation 60.0 72.0 65.4 CNN_PW_Relation 75.6 70.0 72.5 Notebook CNN_PW_R_Flat 67.3 72.0 69.5 Book CNN_PW_R_Flat 75.4 81.0 77.9 CNN_PW_R_Subtree 67.8 76.0 71.6 CNN_PW_R_Subtree 75.2 79.0 76.9 Feature-Based 59.0 81.2 68.4 Feature-Based 66.5 89.1 76.2 Path-Rule-Based (*) 74.1 56.8 64.3 CNN_PathWord 78.0 70.0 73.8 CNN_PathWord 47.9 98.0 64.3 CNN_PW_Relation 73.9 78.0 76.0 CNN_PW_Relation 78.1 71.0 74.2 Hotel CNN_PW_R_Flat 72.1 81.0 76.2 Restaurant CNN_PW_R_Flat 73.5 77.0 75.4 CNN_PW_R_Subtree 77.9 82.0 79.7 CNN_PW_R_Subtree 75.9 79.0 77.3 Feature-Based 61.5 72.1 66.4 Feature-Based 66.9 78.6 72.3 CNN_PathWord 65.8 59.4 62.4 CNN_PW_Relation 71.3 75.3 73.2 All CNN_PW_R_Flat 71.8 78.2 74.9 CNN_PW_R_Subtree 74.2 78.6 76.4 Feature-Based 59.3 79.3 67.9

a) The results indicated by * are obtained from the work of Zhao et al. [8].

Citations

• #### 0

Altmetric

Copyright 2020 Science China Press Co., Ltd. 《中国科学》杂志社有限责任公司 版权所有