跳到主要內容

臺灣博碩士論文加值系統

(44.220.44.148) 您好!臺灣時間:2024/06/21 15:44
字體大小: 字級放大   字級縮小   預設字形  
回查詢結果 :::

詳目顯示

我願授權國圖
: 
twitterline
研究生:林淑鶯
研究生(外文):Shuying Lin
論文名稱:基於深度學習的中文反諷程度預測
論文名稱(外文):Predicting Chinese Irony Intensity Using Deep Learning
指導教授:賴國華禹良治禹良治引用關係
指導教授(外文):K. Robert LaiLiang-Chih Yu
口試委員:常安
口試委員(外文):An Chang
口試日期:2017年7月4日
學位類別:碩士
校院名稱:元智大學
系所名稱:資訊工程學系
學門:工程學門
學類:電資工程學類
論文種類:學術論文
論文出版年:2017
畢業學年度:106
語文別:中文
論文頁數:31
中文關鍵詞:情感分析深度學習中文反諷語料庫反諷程度預測多維度關係模型
外文關鍵詞:Sentiment AnalysisDeep LearningChinese Irony corpusIrony Intensity predictionMulti-dimensional relational model
相關次數:
  • 被引用被引用:1
  • 點閱點閱:437
  • 評分評分:
  • 下載下載:0
  • 收藏至我的研究室書目清單書目收藏:2
在文本情感分析中,對文本進行處理並分析出情感傾向主要有兩類研究,分別是類別性研究和維度型研究,維度型研究體現了更加細細微性的情感資訊。目前中文維度型語料少,維度型研究預測結果準確性低,因此中文維度型研究很難。再加上反諷語料的結構性特殊,導致中文反諷程度預測更難,於是本文針對這些問題,提出了基於深度學習方法的多維度關係模型和線性調整模型,在建立的中文反諷語料庫Valence-Arousal-Irony三維空間中,進行中文反諷程度預測研究。使用深度學習神經網路模型來生成文本的重要特徵句向量,然後利用我們的多維度關係模型和線性調整模型來優化預測結果。多維度關係模型是結合多維度的特徵向量來同時預測VAI的值。多維度關係調整模型是單獨預測結果,然後對預測的結果考慮其他維度的預測結果進行一定的調整。最後我們在NTU反諷語料庫下進行了一系列的實驗,實驗結果表明多維度關係模型和多維度關係調整模型都有效的提高了中文反諷程度的預測結果。
In the Sentiment analysis of text, there are two main types of research on the processing of the text and the analysis of the tendency of sentiment. It is Categorical Sentiment analysis and Dimensional Sentiment analysis respectively. The study of Dimensional Sentiment analysis can provide more fine-grained sentiment information. At present, the lack of Chinese dimensional affective corpus and the low accuracy of dimensional sentiment analysis, which make difficulties in applying dimensional approach for Chinese text. Similarly, it is hard to apply it to predict Chinese Irony intensity because of the speciality of the structure in Irony expression. In order to solve the problem, we propose a multi-dimensional relational model and dimensional relation adjustment model based on the depth learning method. In the establishment of a Chinese Irony corpus, we carry out the study of Chinese Irony intensity prediction in Valence-Arousal-Irony three-dimensional space. We use the deep learning neural network model to generate the most important feature vector of the text, and then use our multi-dimensional relational model and dimensional relation adjustment model to optimize the prediction results. The multi-dimensional relational model is combined with multi-dimensions feature vector for VAI prediction simultaneously. The dimensional relation adjustment model is used to optimize the VAI prediction after predicting the VAI ratings of texts respectively. Finally, we have carried out a series of experiments under NTU irony corpus. The experimental results show that the model can show good performance for Chinese Irony intensity prediction.
摘要 iii
ABSTRACT iv
致謝 v
目錄 vi
表目錄 viii
圖目錄 ix
第一章 引言 1
1.1動機 1
1.2挑戰 2
1.3貢獻 3
1.4組織結構 3
第二章 背景和相關工作 4
2.1詞嵌入 4
2.2文本情感分析 5
2.3文本反諷研究 6
2.3.1 反諷分類 7
2.3.1 反諷維度分析 9
2.3.3 反諷語料 9
2.4深度學習演算法 10
2.4.1 多層感知器 10
2.4.2 卷積神經網路 11
2.4.3 長短時記憶網路 13
第三章 VAI關係模型 15
3.1深度學習模型 15
3.1.1 CNN模型 16
3.1.2 LSTM模型 17
3.1.3 CNN-LSTM模型 18
3.1.4 LSTM-CNN模型 18
3.2 多維度關聯分析 19
3.3 多維度關係模型 22
3.4 多維度關係調整模型 23
第四章 實驗內容 25
4.1實驗資料 25
4.2實驗流程 25
4.3實驗模型說明 26
4.4評估方法 27
4.5實驗結果及分析 27
第五章 總結 31
參考資料 32
[1] Roberts R M, Kreuz R J. Why do people use figurative language?[J]. Psychological science, 1994, 5(3): 159-163.
[2] Bogel F V. Irony, Inference, and Critical Understanding[J]. Yale Review, 503-19.
[3] Tang Y, Chen H H. Chinese Irony Corpus Construction and Ironic Structure Analysis[C]//COLING. 2014: 1269-1278.
[4] Hinton G E. Learning distributed representations of concepts[C]//Proceedings of the eighth annual conference of the cognitive science society. 1986, 1: 12.
[5] Mikolov T, Chen K, Corrado G, et al. Efficient estimation of word representations in vector space[J]. arXiv preprint arXiv:1301.3781, 2013.
[6] Mikolov T, Sutskever I, Chen K, et al. Distributed representations of words and phrases and their compositionality[C]//Advances in neural information processing systems. 2013: 3111-3119.
[7] Pennington J, Socher R, Manning C D. Glove: Global vectors for word representation[C]//EMNLP. 2014, 14: 1532-1543.
[8] Pang B, Lee L. Opinion mining and sentiment analysis[J]. Foundations and Trends® in Information Retrieval, 2008, 2(1–2): 1-135.
[9] Calvo R A, D'Mello S. Affect detection: An interdisciplinary review of models, methods, and their applications[J]. IEEE Transactions on affective computing, 2010, 1(1): 18-37.
[10] Liu B. Sentiment analysis and opinion mining[J]. Synthesis lectures on human language technologies, 2012, 5(1): 1-167.
[11] Feldman R. Techniques and applications for sentiment analysis[J]. Communications of the ACM, 2013, 56(4): 82-89.
[12] Yu L C, Lee L H, Hao S, et al. Building Chinese Affective Resources in Valence-Arousal Dimensions[C]//HLT-NAACL. 2016: 540-545.
[13] Ekman P. An argument for basic emotions[J]. Cognition & emotion, 1992, 6(3-4): 169-200.
[14] Li J, Ott M, Cardie C, et al. Towards a General Rule for Identifying Deceptive Opinion Spam[C]//ACL (1). 2014: 1566-1576.
[15] Baccianella S, Esuli A, Sebastiani F. SentiWordNet 3.0: An Enhanced Lexical Resource for Sentiment Analysis and Opinion Mining[C]//LREC. 2010, 10: 2200-2204.
[16] Ressel J A. A circumplex model of affect[J]. J. Personality and Social Psychology, 1980, 39: 1161-78.
[17] Davidov D, Tsur O, Rappoport A. Semi-supervised recognition of sarcastic sentences in twitter and amazon[C]//Proceedings of the fourteenth conference on computational natural language learning. Association for Computational Linguistics, 2010: 107-116.
[18] Yu L C, Wang J, Lai K R, et al. Predicting Valence-Arousal Ratings of Words Using a Weighted Graph Method[C]//ACL (2). 2015: 788-793.
[19] Wang J, Yu L C, Lai K R, et al. Dimensional sentiment analysis using a regional CNN-LSTM model[C]//ACL 2016—Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics. Berlin, Germany. 2016, 2: 225-230.
[20] Pennebaker J W, Francis M E, Booth R J. Linguistic inquiry and word count: LIWC 2001[J]. Mahway: Lawrence Erlbaum Associates, 2001, 71(2001): 2001.
[21] Bradley M M, Lang P J. Affective norms for English words (ANEW): Instruction manual and affective ratings[R]. Technical report C-1, the center for research in psychophysiology, University of Florida, 1999.
[22] Nielsen F Å. A new ANEW: Evaluation of a word list for sentiment analysis in microblogs[J]. arXiv preprint arXiv:1103.2903, 2011.
[23] Stone P J, Dunphy D C, Smith M S. The general inquirer: A computer approach to content analysis[J]. 1966.
[24] Bradley M M, Lang P J. Affective Norms for English Text (ANET): Affective ratings of text and instruction manual[J]. Techical Report. D-1, University of Florida, Gainesville, FL, 2007.
[25] Huang C L, Chung C K, Hui N, et al. The development of the Chinese linguistic inquiry and word count dictionary[J]. Chinese Journal of Psychology, 2012, 54(2): 185-201.
[26] Ku L W, Chen H H. Mining opinions from the Web: Beyond relevance retrieval[J]. Journal of the Association for Information Science and Technology, 2007, 58(12): 1838-1850.
[27] Wilson D, Sperber D. On verbal irony[J]. Lingua, 1992, 87(1-2): 53-76.
[28] Utsumi A. A unified theory of irony and its computational formalization[C]//Proceedings of the 16th conference on Computational linguistics-Volume 2. Association for Computational Linguistics, 1996: 962-967.
[29] Zhang M, Zhang Y, Fu G. Tweet Sarcasm Detection Using Deep Neural Network[C]//COLING. 2016: 2449-2460
[30]. Hao Y, Veale T. Support structures for linguistic creativity: A computational analysis of creative irony in similes[C]//Proceedings of the Cognitive Science Society. 2009, 31(31).
[31] Veale T, Hao Y. Detecting Ironic Intent in Creative Comparisons[C]//ECAI. 2010, 215: 765-770.
[32] Hao Y, Veale T. An ironic fist in a velvet glove: Creative mis-representation in the construction of ironic similes[J]. Minds and Machines, 2010, 20(4): 635-650.
[33] Carvalho P, Sarmento L, Silva M J, et al. Clues for detecting irony in user-generated contents: oh...!! it's so easy;-[C]//Proceedings of the 1st international CIKM workshop on Topic-sentiment analysis for mass opinion. ACM, 2009: 53-56.
[34] Davidov D, Tsur O, Rappoport A. „A Great Catchy Name: Semi-Supervised Recognition of Sarcastic Sentences in Online Product Reviews “[C]//Proceedings of the Fourth International AAAI Conference on Weblogs and Social Media (ICWSM-2010),(Stroudsburg, PA, USA: ACL, 2010). 2010: 107-116.
[35] Reyes A, Rosso P, Veale T. A multidimensional approach for detecting irony in twitter[J]. Language resources and evaluation, 2013, 47(1): 239-268.
[36] Maynard D, Greenwood M A. Who cares about Sarcastic Tweets? Investigating the Impact of Sarcasm on Sentiment Analysis[C]//LREC. 2014: 4238-4243.
[37] González-Ibánez R, Muresan S, Wacholder N. Identifying sarcasm in Twitter: a closer look[C]//Proceedings of the 49th Annual Meeting of the Association for Computational Linguistics: Human Language Technologies: Short Papers-Volume 2. Association for Computational Linguistics, 2011: 581-586.
[38] Reyes A, Rosso P, Buscaldi D. From humor recognition to irony detection: The figurative language of social media[J]. Data & Knowledge Engineering, 2012, 74: 1-12.
[39] Ptácek T, Habernal I, Hong J. Sarcasm Detection on Czech and English Twitter[C]//COLING. 2014: 213-223.
[41] Riloff E, Qadir A, Surve P, et al. Sarcasm as Contrast between a Positive Sentiment and Negative Situation[C]//EMNLP. 2013, 13: 704-714.
[42] Joshi A, Sharma V, Bhattacharyya P. Harnessing Context Incongruity for Sarcasm Detection[C]//ACL (2). 2015: 757-762.
[43] Sivic J, Zisserman A. Efficient visual search of videos cast as text retrieval[J]. IEEE transactions on pattern analysis and machine intelligence, 2009, 31(4): 591-606.
[44] Buschmeier K, Cimiano P, Klinger R. An Impact Analysis of Features in a Classification Approach to Irony Detection in Product Reviews[C]//WASSA@ ACL. 2014: 42-49.
[45] Wallace B C, Do Kook Choe L K, Kertz L, et al. Humans Require Context to Infer Ironic Intent (so Computers Probably do, too)[C]//ACL (2). 2014: 512-516.
[46] Wallace B C, Do Kook Choe, Charniak E. Sparse, Contextually Informed Models for Irony Detection: Exploiting User Communities, Entities and Sentiment[C]//ACL (1). 2015: 1035-1044.
[47] Khattri A, Joshi A, Bhattacharyya P, et al. Your Sentiment Precedes You: Using an author's historical tweets to predict sarcasm[C]//WASSA@ EMNLP. 2015: 25-30.
[48] Bamman D, Smith N A. Contextualized Sarcasm Detection on Twitter[C]//ICWSM. 2015: 574-577.
[49] Rajadesingan A, Zafarani R, Liu H. Sarcasm detection on twitter: A behavioral modeling approach[C]//Proceedings of the Eighth ACM International Conference on Web Search and Data Mining. ACM, 2015: 97-106.
[50] Ghosh D, Guo W, Muresan S. Sarcastic or Not: Word Embeddings to Predict the Literal or Sarcastic Meaning of Words[C]//EMNLP. 2015: 1003-1012.
[51] Tungthamthiti P, Santus E, Xu H, et al. Sentiment Analyzer with Rich Features for Ironic and Sarcastic Tweets[C]//PACLIC. 2015.
[52] Xu H, Santus E, Laszlo A, et al. LLT-PolyU: Identifying Sentiment Intensity in Ironic Tweets[C]//SemEval@ NAACL-HLT. 2015: 673-678.
[53] Ghosh A, Li G, Veale T, et al. SemEval-2015 Task 11: Sentiment Analysis of Figurative Language in Twitter[C]//SemEval@ NAACL-HLT. 2015: 470-478.
[54] Karoui J, Benamara F, Moriceau V, et al. Towards a contextual pragmatic model to detect irony in tweets[C]//53rd Annual Meeting of the Association for Computational Linguistics (ACL 2015). 2015: PP. 644-650.
[55] Barbieri F, Ronzano F, Saggion H. How Topic Biases Your Results? A Case Study of Sentiment Analysis and Irony Detection in Italian[C]//RANLP. 2015: 41-47.
[56] Filatova E. Irony and Sarcasm: Corpus Generation and Analysis Using Crowdsourcing[C]//LREC. 2012: 392-398.
[57] Reyes A, Rosso P. Mining subjective knowledge from customer reviews: A specific case of irony detection[C]//Proceedings of the 2nd workshop on computational approaches to subjectivity and sentiment analysis. Association for Computational Linguistics, 2011: 118-124.
[58] Swanson R, Lukin S M, Eisenberg L, et al. Getting Reliable Annotations for Sarcasm in Online Dialogues[C]//LREC. 2014: 4250-4257.
[59] Deng L, Yu D. Deep learning: methods and applications[J]. Foundations and Trends® in Signal Processing, 2014, 7(3–4): 197-387.
[60] Kim Y. Convolutional neural networks for sentence classification[J]. arXiv preprint arXiv:1408.5882, 2014.
[61] Iyyer M, Manjunatha V, Boyd-Graber J L, et al. Deep Unordered Composition Rivals Syntactic Methods for Text Classification[C]//ACL (1). 2015: 1681-1691.
[62] Graves A. Supervised sequence labelling with recurrent neural networks[M]. Heidelberg: Springer, 2012.
[63] Wang X, Liu Y, Sun C, et al. Predicting Polarities of Tweets by Composing Word Embeddings with Long Short-Term Memory[C]//ACL (1). 2015: 1343-1353.
[64] Liu P, Joty S R, Meng H M. Fine-grained Opinion Mining with Recurrent Neural Networks and Word Embeddings[C]//EMNLP. 2015: 1433-1443.
[65] Joshi A, Tripathi V, Patel K, et al. Are Word Embedding-based Features Useful for Sarcasm Detection?[J]. arXiv preprint arXiv:1610.00883, 2016.
[66] Ghosh A, Veale T. Fracking Sarcasm using Neural Network[C]//WASSA@ NAACL-HLT. 2016: 161-169.
[67] Amir S, Wallace B C, Lyu H, et al. Modelling context with user embeddings for sarcasm detection in social media[J]. arXiv preprint arXiv:1607.00976, 2016.
[68] Yao Z, Wu J, Zhang Y, et al. Norms of valence, arousal, concreteness, familiarity, imageability, and context availability for 1,100 Chinese words[J]. Behavior research methods, 2016: 1-12.
[69] Lang P J. Behavioral treatment and bio-behavioral assessment: Computer applications[J]. 1980.
電子全文 電子全文(本篇電子全文限研究生所屬學校校內系統及IP範圍內開放)
連結至畢業學校之論文網頁點我開啟連結
註: 此連結為研究生畢業學校所提供,不一定有電子全文可供下載,若連結有誤,請點選上方之〝勘誤回報〞功能,我們會盡快修正,謝謝!
QRCODE
 
 
 
 
 
                                                                                                                                                                                                                                                                                                                                                                                                               
第一頁 上一頁 下一頁 最後一頁 top
無相關期刊