Abstract
Sentiment analysis in Danmaku video interaction aims at measuring public mood in respect of the video, which is helpful for the potential applications in behavioral science. Once these sentiments are discovered, this feedback can help video creators improve the video quality and greatly enhance online users’ watching experience. Predicting these entity-level sentiments is challenging because there is no publicly available dataset about entity-level sentiment analysis of Danmaku-enabled video comments. Furthermore, the targeted entity with skewed unbalance distribution in real-world scenarios, making the task more challenging, especially when the target entity only has positive (negative) emotional comments. In this case, applying previous aspect-level sentiment analysis models directly will introduce entity bias. In this paper, we propose a large-scale Chinese video comments dataset containing time-sync Danmaku comments and traditional video comments, targeting multiple entities and sentiments associated with each entity from popular video websites. We also propose a framework of entity-level sentiment analysis with two de-biasing models: hard-masking de-bias model and soft-masking de-bias model. This framework is defined by parallel neural networks to learn the representation of comments sentences. Based on the representations, our model learns a masking strategy for entity words to avoid overfitting and mitigate the bias. Our experiments on Danmaku-enabled video datasets show that the soft-masking model significantly outperforms comparable baselines, with a relative F1-score improvement of 9.33% compared to AEN-BERT and a relative F1-score improvement of 45.77% compared to Td-LSTM. Furthermore, experiments on different distribution bias of entity demonstrate that our proposed model can achieve competitive performances. The findings of this research have implications for measuring public sentiment for entities mentioned in a specific video domain. It can also be used as a benchmark dataset for aspect entity sentiment detection methods.
Similar content being viewed by others
Notes
References
Alarifi A, Tolba A, Al-Makhadmeh Z, Said W (2020) A big data approach to sentiment analysis using greedy feature selection with cat swarm optimization-based long short-term memory neural networks. J Supercomput 76(6):4414–4429. https://doi.org/10.1007/s11227-018-2398-2
Bai Q, Hu Q, Fang F, He L (2018) Topic detection with danmaku: a time-sync joint NMF approach. DEXA 11030:428–435
Cai Y, Wan X (2019a) Multi-domain sentiment classification based on domain-aware embedding and attention. In: IJCAI-19, IJCAI, pp 4904–4910
Cai Y, Wan X (2019b) Multi-domain sentiment classification based on domain-aware embedding and attention. In: Kraus S (ed) IJCAI. ijcai.org, pp 4904–4910
Cao Y, Xu H (2020) Satnet: Symmetric adversarial transfer network based on two-level alignment strategy towards cross-domain sentiment classification (student abstract). In: AAAI, pp 13763–13764
Chen Y, Gao Q, Rau PL (2017b) Watching a movie alone yet together: understanding reasons for watching Danmaku videos. In: International Journal of Human–Computer Interaction
Chen X, Zhang Y, Ai Q, Xu H, Yan J, Qin Z (2017a) Personalized key frame recommendation. In: ACM SIGIR, 2017. ACM, pp 315–324
Cho K, van Merrienboer B, Gülçehre Ç, Bahdanau D, Bougares F, Schwenk H, Bengio Y (2014) Learning phrase representations using RNN encoder–decoder for statistical machine translation. In: EMNLP, pp 1724–1734
Cui Y, Che W, Liu T, Qin B, Yang Z, Wang S, Hu G (2019) Pre-training with whole word masking for chinese bert. arXiv preprint arXiv:190608101
Devlin J, Chang MW, Lee K, Toutanova K (2018) Bert: pre-training of deep bidirectional transformers for language understanding. arXiv preprint arXiv:181004805
Dong L, Wei F, Tan C, Tang D, Zhou M, Xu K (2014) Adaptive recursive neural network for target-dependent twitter sentiment classification. In: Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics, ACL 2014, June 22–27, 2014, Baltimore, MD, USA, Volume 2: Short Papers. The Association for Computer Linguistics, pp 49–54. https://doi.org/10.3115/v1/p14-2009
Felsenthal DS, Machover M (2001) The treaty of nice and qualified majority voting. Soc Choice Welf 18(3):431–464
Gonen H, Goldberg Y (2019) Lipstick on a pig: Debiasing methods cover up systematic gender biases in word embeddings but do not remove them. In: NAACL HLT 2019–2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies—Proceedings of the Conference, pp 609–614
He M, Ge Y, Wu L, Chen E, Tan C (2016) Predicting the popularity of DanMu—enabled videos: a multi-factor view. Springer, Berlin
Khot T, Clark P, Guerquin M, Jansen P, Sabharwal A (2020) QASC: a dataset for question answering via sentence composition. In: The Thirty-Fourth AAAI Conference on Artificial Intelligence, AAAI 2020, The Thirty-Second Innovative Applications of Artificial Intelligence Conference, IAAI 2020, the Tenth AAAI Symposium on Educational Advances in Artificial Intelligence, EAAI 2020, New York, NY, USA, February 7–12, pp 8082–8090
Kingma DP, Ba J (2015) Adam: a method for stochastic optimization. In: ICLR 2015
Kurita K, Vyas N, Pareek A, Black AW, Tsvetkov Y (2019) Measuring bias in contextualized word representations. CoRR abs/1906.07337. arxiv:1906.07337
Lei J, Zhang Q, Wang J, Luo H (2019) BERT based hierarchical sequence classification for context-aware microblog sentiment analysis. In: Gedeon T, Wong KW, Lee M (eds) ICONIP, Lecture Notes in Computer Science, vol 11955, pp 376–386
Li X, Bing L, Lam W, Shi B (2018) Transformation networks for target-oriented sentiment classification. arXiv preprint. arXiv:180501086
Lin C, Zhao S, Meng L, Chua T (2020) Multi-source domain adaptation for visual sentiment classification. In: AAAI. AAAI Press, pp 2661–2668
Lv G, Tong X, Chen E, Yi Z, Yi Z (2016) Reading the videos: temporal labeling for crowdsourced time-sync videos based on semantic embedding. In: AAAI, pp 3000–3006
Maas AL, Daly RE, Pham PT, Huang D, Potts C (2011) Learning word vectors for sentiment analysis. In: Meeting of the Association for Computational Linguistics: Human Language Technologies
Ma X, Cao N (2017) Video-based evanescent, anonymous, asynchronous social interaction: motivation and adaption to medium. In: ACM CSCW, pp 770–782
Ma S, Cui L, Dai D, Wei F, Sun X (2019) Livebot: generating live video comments based on visual and textual contexts. In: AAAI
Maudslay RH, Gonen H, Cotterell R, Teufel S (2019) It’s all in the name: mitigating gender bias with name-based counterfactual data substitution. EMNLP
Maudslay RH, Gonen H, Cotterell R, Teufel S (2020) It’s all in the name: mitigating gender bias with name-based counterfactual data substitution. EMNLP, pp 5267–5275
Mehrabi N, Morstatter F, Saxena N, Lerman K, Galstyan A (2019) A survey on bias and fairness in machine learning. CoRR abs/1908.09635. arxiv:1908.09635
Qian X, Liu X, Ma X, Lu D, Xu C (2016) What is happening in the video? Annotate video by sentence. IEEE Trans Circuits Syst Video Technol 26(9):1746–1757
Ruan S, Zhang K, Wang Y, Tao H, He W, Lv G, Chen E (2020) Context-aware generation-based net for multi-label visual emotion recognition. In: ICME, pp 1–6
Saeidi M, Bouchard G, Liakata M, Riedel S (2016) Sentihood: Targeted aspect based sentiment analysis dataset for urban neighbourhoods. In: Calzolari N, Matsumoto Y, Prasad R (eds) COLING 2016, 26th International Conference on Computational Linguistics, Proceedings of the Conference: Technical Papers, December 11–16, 2016, Osaka, Japan, ACL, pp 1546–1556. https://www.aclweb.org/anthology/C16-1146/
Song Y, Shi S, Li J, Zhang H (2018) Directional skip-gram: explicitly distinguishing left and right context for word embeddings. In: ACL, pp 175–180
Song Y, Wang J, Jiang T, Liu Z, Rao Y (2019) Attentional encoder network for targeted sentiment classification. arXiv preprint. arXiv:190209314
Su Y, Hu W, Jiang J, Su R (2020) A novel LMAEB-CNN model for Chinese microblog sentiment analysis. J Supercomput 76(11):9127–9141. https://doi.org/10.1007/s11227-020-03198-x
Tang D, Qin B, Feng X, Liu T (2015) Effective lstms for target-dependent sentiment classification. arXiv preprint. arXiv:151201100
Tang D, Qin B, Liu T (2016) Aspect level sentiment classification with deep memory network. arXiv preprint arXiv:160508900
Wang B, Yao T, Zhang Q, Xu J, Wang X (2020) Reco: A large scale Chinese reading comprehension dataset on opinion. In: AAAI, pp 9146–9153
Xian Y, Li J, Zhang C, Liao Z (2015) Video highlight shot extraction with time-sync comment. In: International Workshop on Hot Topics in Planet-Scale Mobile Computing and Online Social Networking, pp 31–36
Yang X, Binglu W, Junjie H, Shuwen L (2017b) Natural language processing in “bullet screen” application. In: ICSSSM. IEEE, pp 1–6
Yang W, Ruan N, Gao W, Wang K, Ran W, Jia W (2017a) Crowdsourced time-sync video tagging using semantic association graph. In: ICME, 2017. IEEE, pp 547–552
Yao Y, Bort J, Huang Y (2017) Understanding Danmaku’s potential in online video learning. CHI 2017:3034–3040
Yu J, Jiang J (2019) Adapting bert for target-oriented multimodal sentiment classification. In: Proceedings of the Twenty-Eighth International Joint Conference on Artificial Intelligence, IJCAI-19, IJCAI, pp 5408–5414
Zeng B, Yang H, Xu R, Zhou W, Han X (2019) LCF: A Local context focus mechanism for aspect-based sentiment classification. Applied Sciences (Switzerland) 9(16), https://doi.org/10.3390/app9163389
Zhao Y, Peng X, Tang J, Song S (2017) Understanding young people’s we-intention to contribute in Danmaku websites: motivational, social, and subculture influence. In: Conference 2017 Proceedings
Zhou J, Chen Q, Huang JX, Hu QV, He L (2020a) Position-aware hierarchical transfer model for aspect-level sentiment classification. Inf Sci 513:1–16
Zhou J, Huang JX, Hu QV, He L (2020b) Is position important? Deep multi-task learning for aspect-based sentiment analysis. Appl Intell 50(10):3367–3378
Zhou J, Huang JX, Hu QV, He L (2020d) SK-GCN: modeling syntax and knowledge via graph convolutional network for aspect-level sentiment classification. Knowl Based Syst 205:106292
Zhou J, Huang JX, Hu QV, He L (2020c) Modeling multi-aspect relationship with joint learning for aspect-level sentiment classification. In: DASFAA, pp 786–802
Zhou J, Tian J, Wang R, Wu Y, Xiao W, He L (2020e) Sentix: A sentiment-aware pre-trained model for cross-domain sentiment analysis. In: Proceedings of the 28th International Conference on Computational Linguistics, COLING 2020, Barcelona, Spain (Online), December 8–13
Acknowledgements
The computation is performed in ECNU Multifunctional Platform for Innovation (001). The authors would like to thank Shijun Zhang and Kai Song for excellent technical support and data annotation support. The authors would also like to thank the editors and the reviewers for their high-quality and constructive comments. These suggestions and comments are important and useful.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
About this article
Cite this article
Bai, Q., Wei, K., Zhou, J. et al. Entity-level sentiment prediction in Danmaku video interaction. J Supercomput 77, 9474–9493 (2021). https://doi.org/10.1007/s11227-021-03652-4
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11227-021-03652-4