›› 2017, Vol. 32 ›› Issue (4): 785-795.doi: 10.1007/s11390-017-1759-2

Special Issue: Artificial Intelligence and Pattern Recognition

• Special Issue on Deep Learning • Previous Articles     Next Articles

Emphasizing Essential Words for Sentiment Classification Based on Recurrent Neural Networks

Fei Hu1,2, Student Member, CCF, Li Li1,*, Senior Member, CCF, Member, ACM, Zi-Li Zhang1, Distinguished Member, CCF, Member, ACM, Jing-Yuan Wang1, Student Member, CCF, Xiao-Fei Xu1, Student Member, CCF   

  1. 1 College of Computer and Information Science, Southwest University, Chongqing 400715, China;
    2 Network Centre, Chongqing University of Education, Chongqing 400065, China
  • Received:2016-12-20 Revised:2017-06-03 Online:2017-07-05 Published:2017-07-05
  • Contact: Li Li E-mail:lily@swu.edu.cn
  • Supported by:

    The work was supported by the Scientific and Technological Research Program of Chongqing Municipal Education Commission of China under Grant No. KJ1501405, the National Natural Science Foundation of China under Grant No. 61170192, and the Chongqing Science and Technology Commission (CSTC) under Grant No. cstc2015gjhz40002.

With the explosion of online communication and publication, texts become obtainable via forums, chat messages, blogs, book reviews and movie reviews. Usually, these texts are much short and noisy without sufficient statistical signals and enough information for a good semantic analysis. Traditional natural language processing methods such as Bow-of-Word (BOW) based probabilistic latent semantic models fail to achieve high performance due to the short text environment. Recent researches have focused on the correlations between words, i.e., term dependencies, which could be helpful for mining latent semantics hidden in short texts and help people to understand them. Long short-term memory (LSTM) network can capture term dependencies and is able to remember the information for long periods of time. LSTM has been widely used and has obtained promising results in variants of problems of understanding latent semantics of texts. At the same time, by analyzing the texts, we find that a number of keywords contribute greatly to the semantics of the texts. In this paper, we establish a keyword vocabulary and propose an LSTM-based model that is sensitive to the words in the vocabulary; hence, the keywords leverage the semantics of the full document. The proposed model is evaluated in a short-text sentiment analysis task on two datasets:IMDB and SemEval-2016, respectively. Experimental results demonstrate that our model outperforms the baseline LSTM by 1% 2% in terms of accuracy and is effective with significant performance enhancement over several non-recurrent neural network latent semantic models (especially in dealing with short texts). We also incorporate the idea into a variant of LSTM named the gated recurrent unit (GRU) model and achieve good performance, which proves that our method is general enough to improve different deep learning models.

[1] Wang G, Zhang Z, Sun J S, Sun J S, Yang S L, Larsonc C A. POS-RS:A random subspace method for sentiment classification based on part-of-speech analysis. Information Processing & Management, 2015, 51(4):458-479.

[2] Hua W, Wang Z Y, Wang H X, Zheng K, Zhou X F. Short text understanding through lexical-semantic analysis. In Proc. Int. Conf. Data Engineering, April 2015, pp.495-506.

[3] Zou H, Tang X H, Xie B, Liu B. Sentiment classification using machine learning techniques with syntax features. In Proc. Int. Conf. Computational Science and Computational Intelligence, Dec. 2015, pp.175-179.

[4] Davuth N, Kim S R. Classification of malicious domain names using support vector machine and bi-gram method. International Journal of Security and its Applications, 2013, 7(1):51-58.

[5] Bao S H, Xu S L, Zhang L, Yan R, Su Z, Han D Y, Yu Y. Mining social emotions from affective text. IEEE Trans. Knowledge and Data Engineering, 2012, 24(9):1658-1670.

[6] Rao Y H, Lei J S, Liu W Y, Li Q, Chen M L. Building emotional dictionary for sentiment analysis of online news. World Wide Web, 2014, 17(4):723-742.

[7] Stoyanov V, Cardie C. Annotating topics of opinions. In Proc. the 6th International Conference on Language Resources and Evaluation, May 31-June 1, 2008, pp.3213-3217.

[8] Cheng X Q, Yan X H, Guo Y Y, Guo J F. BTM:Topic modeling over short texts. IEEE Trans. Knowledge and Data Engineering, 2014, 26(12):2928-2941.

[9] Wang Z Y, Zhao K J, Wang H X, Meng X F, Wen J R. Query understanding through knowledge-based conceptualization. In Proc. the 24th Int. Conf. Artificial Intelligence, July 2015, pp.3264-3270.

[10] Cheng J P, Wang Z Y, Wen J R, Yan J. Contextual text understanding in distributional semantic space. In Proc. the 24th ACM Int. Conf. Information and Knowledge Management, Oct. 2015, pp.133-142.

[11] Cui W Y, Zhou X Y, Lin H Y, Xiao Y H. Verb pattern:A probabilistic semantic representation on verbs. In Proc. the 30th AAAI Conf. Artificial Intelligence, March 2016, pp.2587-2593.

[12] Zhang X W, Wu B. Short text classification based on feature extension using the n-gram model. In Proc. the 12th Int. Conf. Fuzzy Systems and Knowledge Discovery, Aug. 2015, pp.710-716.

[13] López G J, Ruiz I M. Character and word baselines systems for irony detection in Spanish short texts. Procesamiento de Lenguaje Natural, 2016, 56:41-48.

[14] Song G, Ye Y M, Du X L, Huang X H, Bie S F. Short text classification:A survey. Journal of Multimedia, 2014, 9(5):635-643.

[15] Wang M, Lin L F, Wang F. Improving short text classification through better feature space selection. In Proc. the 9th Int. Conf. Computational Intelligence and Security, December 2013, pp.120-124.

[16] Wang B K, Huang Y F, Yang W X, Li X. Short text classification based on strong feature thesaurus. Journal of Zhejiang University Science C, 2012, 13(9):649-659.

[17] Kim K, Chung B S, Choi Y, Lee S, Jung J Y, Park J. Language independent semantic kernels for short-text classification. Expert Systems with Applications, 2014, 41(2):735-743.

[18] Fan X H, Hu H G. Construction of high-quality feature extension mode library for Chinese short-text classification. In Proc. WASE Int. Conf. Information Engineering, Aug. 2010, pp.87-90.

[19] Song Y Q, Wang H X, Wang Z Y, Li H S, Chen W Z. Short text conceptualization using a probabilistic knowledgebase. In Proc. the 22nd Int. Joint Conf. Artificial Intelligence, July 2011, pp.2330-2336.

[20] Kim D, Wang H X, Oh A. Context-dependent conceptualization. In Proc. the 23rd Int. Joint Conf. Artificial Intelligence, Aug. 2013, pp.2654-2661.

[21] Huang P S, He X D, Gao J F, Deng L, Acero A, Heck L. Learning deep structured semantic models for web search using clickthrough data. In Proc. the 22nd ACM Int. Conf. Information & Knowledge Management, Oct. 2013, pp.2333-2338.

[22] Shen Y L, He X D, Gao J F, Deng L, Mesnil G. A latent semantic model with convolutional-pooling structure for information retrieval. In Proc. the 23rd ACM Int. Conf. Information and Knowledge Management, Nov. 2014, pp.101-110.

[23] Hochreiter S, Schmidhuber J. Long short-term memory. Neural Computation, 1997, 9(8):1735-1780.

[24] Hu F, Xu X F, Wang J Y, Yang Z B, Li L. Memoryenhanced latent semantic model:Short text understanding for sentiment analysis. In Proc. Int. Conf. Database Systems for Advanced Applications. March 2017, pp.393-407.

[25] Hofmann T. Unsupervised learning by probabilistic latent semantic analysis. Machine Learning, 2001, 42(1/2):177-196.

[26] Wang J, Peng J X, Liu O. A classification approach for less popular webpages based on latent semantic analysis and rough set model. Expert Systems with Applications, 2015, 42(1):642-648.

[27] Ke X H, Luo H J. Using LSA and PLSA for text quality analysis. In Proc. Int. Conf. Electronic Science and Automation Control, Jan. 2015, pp.289-291.

[28] Anoop V S, Prem S C, Asharaf S, Alessandro Z. Generating and visualizing topic hierarchies from microblogs:An iterative latent dirichlet allocation approach. In Proc. Int. Conf. Advances in Computing, Communications and Informatics, Aug. 2015, pp.824-828.

[29] Gao J F, Toutanova K, Yih W T. Clickthrough-based latent semantic models for web search. In Proc. the 34th Int. ACM SIGIR Conf. Research and Development in Information Retrieval, July 2011, pp.675-684.

[30] Hinton G E, Salakhutdinov R R. Reducing the dimensionality of data with neural networks. Science, 2006, 313(5786):504-507.

[31] Bengio Y, Ducharme R, Vincent P, Janvin C. A neural probabilistic language model. Journal of Machine Learning Research, 2003, 3(2):1137-1155.

[32] Huang E H, Socher R, Manning C D, Ng A Y. Improving word representations via global context and multiple word prototypes. In Proc. the 50th Annual Meeting of the Association for Computational Linguistics:Long Papers-Volume 1, July 2012, pp.873-882.

[33] Salakhutdinov R, Hinton G. Semantic hashing. International Journal of Approximate Reasoning, 2009, 50(7):969-978.

[34] Mikolov T, Karafiát M, Burget L, ?ernocký J, Khudanpur S. Recurrent neural network based language model. In Proc. the 11th Annual Conference of the International Speech Communication Association, Sept. 2010, 1045-1048.

[35] Mikolov T. Statistical language models based on neural networks. http://www.fit.vutbr.cz/~imikolov/rnnlm/google.pdf, March 2015.

[36] Williams R J, Zipser D. Gradient-based learning algorithms for recurrent networks and their computational complexity. In Backpropagation:Theory, Architectures, and Applications, Chauvin Y, Rumelhart D E (eds.), Lawrence Erlbaum Associates, Inc., 1995, pp.433-486.

[37] Pascanu R, Mikolov T, Bengio Y. On the difficulty of training recurrent neural networks. In Proc. the 30th Int. Conf. Machine Learning, June 2013, pp.1310-1318.

[38] Hochreiter S. The vanishing gradient problem during learning recurrent neural nets and problem solutions. International Journal of Uncertainty, Fuzziness and Knowledge-Based Systems, 1998, 6(2):107-116.

[39] Olah C. Understanding LSTM networks. http://colah.github. io/posts/2015-08-Understanding-LSTMs/, Sept. 2016.

[40] Gers F A, Schmidhuber J, Cummins F. Learning to forget:Continual prediction with LSTM. Neural Computation, 2000, 12(10):2451-2471.

[41] Gers F A, Schmidhuber J. Recurrent nets that time and count. In Proc. the IEEE-INNS-ENNS Int. Joint Conf. Neural Networks, July 2000.

[42] Greff K, Srivastava R K, Koutník J, Steunebrink B R, Schmidhuber J. LSTM:A search space odyssey. IEEE Trans. Neural Networks and Learning Systems, 2015, PP(99):1-11.

[43] Cho K, van Merriënboer B, Gulcehre C, Bahdanau D, Bougares F, Schwenk H, Bengio Y. Learning phrase representations using RNN encoder-decoder for statistical machine translation. arXiv:1406.1078, 2014. http://arxiv. org/abs/1406.1078, Sept. 2016.

[44] Esuli A, Sebastiani F. SentiWordNet:A publicly available lexical resource for opinion mining. In Proc. the 5th Conf. Language Resources and Evaluation, May 2006, pp.417-422.

[45] Baccianella S, Esuli A, Sebastiani F. SentiWordNet 3.0:An enhanced lexical resource for sentiment analysis and opinion mining. In Proc. the 7th Conf. Int. Language Resources and Evaluation, Jan. 2010, pp.2200-2204.

[46] Miller G A. WordNet:A lexical database for English. Communications of the ACM, 1995, 38(11):39-41.

[47] Maas A L, Daly R E, Pham P T, Huang D, Ng A Y, Potts C. Learning word vectors for sentiment analysis. In Proc. the 49th Annual Meeting of the Association for Computational Linguistics:Human Language Technologies-Volume 1, June 2011, pp.142-150.

[48] Nakov P, Ritter A, Rosenthal S, Sebastiani F, Stoyanov V. Evaluation measures for the semeval-2016 task 4:Sentiment analysis in twitter. http://alt.qcri.org/semeval2016/task4/, Feb. 2017.

[49] LeCun Y, Bottou L, Orr G B, Müller K R. Efficient backprop. In Neural Networks:Tricks of the Trade, Orr G B, Müller K R (eds.), Springer, 2012, pp.9-50.

[50] Zeiler M D. ADADELTA:An adaptive learning rate method. arXiv:1212.5701, 2012. https://arxiv.org/abs/1212.5701, Sept. 2016.

[51] Duchi J, Hazan E, Singer Y. Adaptive subgradient methods for online learning and stochastic optimization. Journal of Machine Learning Research, 2011, 12:2121-2159.

[52] Kingma D, Ba J. Adam:A method for stochastic optimization. arXiv:1412.6980, 2014. https://arxiv.org/abs/1412.6980, Sept. 2016.

[53] Graves A, Wayne G, Danihelka I. Neural Turing machines. arXiv:1410.5401, 2014. https://arxiv.org/abs/1410.5401, Sept. 2016.
No related articles found!
Full text



[1] Min Yinghua; Han Zhide;. A Built-in Test Pattern Generator[J]. , 1986, 1(4): 62 -74 .
[2] Wu Yunzeng;. On the Development of Applications of Logic in Programming[J]. , 1987, 2(1): 30 -34 .
[3] Wang Lei; Tan Ying;. The Researches in Fault-Tolerant D ataflow Architecture[J]. , 1991, 6(4): 395 -398 .
[4] Zhang Bo; Zhang Ling;. An Algorithm for Finding D-Time Table[J]. , 1992, 7(1): 62 -67 .
[5] Ye Shiwei; Shi Zhongzhi;. A Necessary Condition about the Optimum Partition on a Finite Set of Samples and Its Application to Clustering Analysis[J]. , 1995, 10(6): 545 -556 .
[6] WU Jie;. Reliable Communication on Cube-Based Multicomputers[J]. , 1996, 11(3): 208 -221 .
[7] Tang Changjie; Xiong Min;. The Temporal Mechanisms in Hbase[J]. , 1996, 11(4): 365 -371 .
[8] Xue Jinyun;. Unified Approach for Developing EfficientAlgorithmic Programs[J]. , 1997, 12(4): 314 -329 .
[9] WANG Deqiang; ZHAO Lianchang;. The Twisted-Cube Connected Networks[J]. , 1999, 14(2): 181 -187 .
[10] WANG Xiaodong; XU Ming; ZHOU Xingming;. Fast Multicast on Multistage Interconnection Networks Using Multi-Head Worms[J]. , 1999, 14(3): 250 -258 .

ISSN 1000-9000(Print)

CN 11-2296/TP

Editorial Board
Author Guidelines
Journal of Computer Science and Technology
Institute of Computing Technology, Chinese Academy of Sciences
P.O. Box 2704, Beijing 100190 P.R. China
E-mail: jcst@ict.ac.cn
  Copyright ©2015 JCST, All Rights Reserved