WANG Ling,TAO Hongcai.Research on the Classification Model of Pre-fusion Chinese Emotion Tendency based on LSTM[J].Journal of Chengdu University of Information Technology,2020,35(02):139-145.[doi:10.16836/j.cnki.jcuit.2020.02.003]
基于LSTM前融合中文情感倾向分类模型的研究
- Title:
- Research on the Classification Model of Pre-fusion Chinese Emotion Tendency based on LSTM
- 文章编号:
- 2096-1618(2020)02-0139-07
- 分类号:
- TP391.1
- 文献标志码:
- A
- 摘要:
- 在互联网平台上,用户可以针对电影、新闻等发表自己的观点、表达自己的情感,为其他用户提供消费该商品的参考意见,也帮助产品经理制定有效的产品消费策略。目前,针对中文情感倾向分类、深度学习的方法取得了一定的成就,尤其是长短期记忆神经网络(LSTM)。该网络是一个时序模型,可以很好地理解评论语义抓住评论中蕴含的情感倾向,但是它存在词向量构建阶段无法突出情感词的情感信息,以及无法针对不同场景进行文本情感倾向分析的问题。为此,提出LSTM前融合情感倾向分类模型。新模型利用情感词的情感标签修正情感词向量,解决了情感词向量无法突出情感信息的问题,并且将电影的简介作为一个输入特征融合到最终句子的特征向量中,实现针对具体的电影新闻场景评论情感倾向分类。实验结果表明,新模型相对于基本的LSTM模型取得了更好的效果,亦表明该模型能更加精确地抓取评论的情感信息。
- Abstract:
- Users can express their own opinions and emotions on the Internet platform for movies, news, etc., provide other users with reference opinions on consumption of the product, and help product strategists to formulate effective product consumption strategies. At present, although for the classification of Chinese affective tendencies, deep learning method has made achievements, especially Long Short-Term Memory Model(LSTM). The network is a time series model, which can well understand the comment semantics and grasp the emotional tendency contained in the comment. However, it has the problems that the emotional information of emotional words cannot be highlighted in the construction stage of word vector and the emotional tendency analysis of text cannot be carried out for different scenes. For this, a LSTM pre-fusion emotional tendency classification model is proposed in this paper, which uses emotional labels of emotional words to modify the emotional words vector, solves the problem that emotional words vector can not highlight emotional information, and integrates the brief introduction of the movie as an input feature into the final sentence feature vector to achieve the emotional tendency classification of specific movie news scene comments. The experimental results show that the novel model is better than the basic LSTM model, and can capture the emotion information of comments more accurately.
参考文献/References:
[1] Liu B.Sentiment analysis:mining opinions,sentiments,and emotions[C].The Cambridge University Press,2015.
[2] Liu B.Sentiment analysis and opinion mining(introduction and survey),Morgan & Claypool,May 2012.
[3] Pang B.,Lee L.A sentimental education:Sentiment analysis using subjectivity summ-arization based on minimum cuts[C].Proc-eedings of the 42nd Annual Meeting on Ass-ociation for Computational Linguistics.Stro-udsburg,PA,USA:ACL,2004:271-278.
[4] 百度百科.文本情感分析[DB/OL].https://baike.baidu.com/item/%E6%96%87%E6%9C%AC%E6%83%85%E6%84%9F%E5%88%86%E6%9E%90/19431243?fr=aladdin,2019-01-12.
[5] Socher R,Perelygin A,Wu J Y,et al.Re-cursive deep models for semantic compositionality over a sentiment treebank[C].Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing,EMNLP.2013:1631-1642.
[6] Ma X L,Jin B Y,Fan B S.An Analysis of Chinese Text Emotional Tendency[J].Information and Documentation Service,2013(1):52-56.
[7] 黄萱菁,张奇,吴苑斌.文本情感倾向分析[J].中文信息学报,2011,25(6):118-126.
[8] 赵妍妍,秦兵,刘挺.文本情感分析[J].软件学报,2010,21(8):1834-1848.
[9] 李艺红,蒋秀凤.中文句子倾向性分析[J].福州大学学报(自然科学版),2010,38(4):504-508.
[10] 夏玉芹,单雪微.基于Python的简单文本情感分析[J].阴三学刊,2018,32(4):58-62.
[11] 周虎,于跃,贾媛媛,等.基于深度LTSM神经网络的在线消费评论情感分类研究[J].中华医学图书情杂志,2018,27(5):23-29.
[12] 黄东晋,纪浩,耿晓云,等.基于文本矢量特征的电影评分预测模型[J].现代电影技术,2019(3):44-50.
[13] 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.
[14] 胡荣磊,芮璐,齐筱,等.基于循环神经网络和注意力模型的文本情感分析[J].计算机应用研究,2019,36(11):3282-3285.
[15] 金宸,李维华,姬晨,等.基于双向LSTM神经网络模型的中文分词[J].中文信息学报,2018,32(2):29-37.
[16] 张玉环.基于多种LSTM结构的文本情感分析[D].北京:北京邮电大学,2018.
[17] Wang J.,Cao Z.W.Chinese Text Sentiment Analysis Using LSTM Network Based on L2 and Nadam [C]. Proceedings of 2017 17th IEEE International Conference on Communication Technology(ICCT 2017),2017:1891-1895.
相似文献/References:
[1]秦雷亮,何培宇,方安成,等.基于CNN-LSTM的柯氏音五时相分类方法研究[J].成都信息工程大学学报,2022,37(02):125.[doi:10.16836/j.cnki.jcuit.2022.02.002]
QIN Leiliang,HE Peiyu,FANG Ancheng,et al.Research on Five-phase Classification Method of Korotkoff Sounds based on CNN-LSTM[J].Journal of Chengdu University of Information Technology,2022,37(02):125.[doi:10.16836/j.cnki.jcuit.2022.02.002]
备注/Memo
收稿日期:2019-12-13 基金项目:国家自然科学基金资助项目(61806170)