本論文的主要目的以應用文件分類的方法來實作以文字為基礎的情緒偵測。在很多過去的情緒相關研究大多注重在語音與臉部表情變化上，對於以文字為基礎的研究則較少著墨。本研究主張將單一句子當作一個極短的文件，以此極短的文件來做情緒偵測。我們的資料來源採用目前網路上非常流行的微型部落格(micro-blog)作為，因為微型部落格的文章字數比較符合單一語句的標準，且使用者會將撰寫文章當時的情緒利用圖示的方式呈現在語句中。我們利用point-wise mutual information 來計算詞與情緒之間的相關強度，選擇強度較高的詞當做候選的特徵詞，並設計一個門檻值，以利挑選品質較好的特徵詞。由於中文特徵詞可能會有同意詞存在，因此再用中文的同意辭林降低特徵空間的維度。最後以「支援向量機(Support Vector Machine)分類演算法」來實作分類的模組。實驗結果顯示，我們所提出的方法可以達到八成以上的辨識率，且發現此分類方法只需要用到很少的特偵數量即可得到不錯的準確率，希望此方法未來可以應用在文件分類的相關研究。

This thesis presents a method of detecting the emotion from the Chinese sentences by using document classification method. There are much work has been done to detect users’ emotion states from multimodal sources such as audio, gestures, and eye gazes over the last decade. But compare to research of emotion detection in multimodal fields, emotion detection from text is still not mature and requires more improvements to be assembled as practical applications. We used the micro-blog as the data source, because the data type of micro-blog is similar to the sentence, and the emotion icon will be annotated when user write the article. The Point-wise Mutual Information (PMI) was employed to help us to calculate the strength between the word and emotion icon. The threshold was designed to avoid choosing the feature which has the similar PMI vale. We reduce the feature dimension by using the Chinese thesaurus. Finally, we use the support vector machine to deal with classification task. The experiments show that the best performance of our method is more than 85%. We find out that the small feature set size was also efficient in the classification task. Our method maybe can be applied to other research of document classification problem.

Contents

Chapter 1 Introduction 1
1.1 Problem Description. 1
1.2 Organization of Thesis. 2
Chapter 2 Related Works 3
2.1 Text based emotion classification. 3
2.2 Support Vector Machine. 6
Chapter 3 Proposed Method 9
3.1 About Corpus and Preprocessing. 10
3.1.1 Micro blog : Plurk.com 10
3.1.2 Preprocessing : CKIP 12
3.2 Feature Selection. 13
3.2.1 Point-wise Mutual Information 13
3.2.2 Feature dimension reduction Using a Chinese Thesaurus 16
3.2.3 Feature Vector Transfer 17
Chapter 4 Experiment 19
4.1 Data, Parameters, and Evaluation Method . 19
4.2 Experiment Results. 20
4.2.1 Only use the top fn PMI score (fn- PMI) 20
4.2.2 PMI score with different thresholds (fn-PMI-T) 21
4.2.3 PMI-T with feature dimension reduction (fn-PMI-T-R) 22
4.3 Misclassification Analysis. 27
Chapter 5 Conclusion and Future work 29
References 31

[1] J. C. Ang, Prosodic cues for emotion recognition in communicator dialogs, M.S. thesis, University of California Berkeley, San Francsico, 2002.
[2] D. B. Bracewell, “Semi-Automatic Creation of an Emotion Dictionary Using WordNet and its Evaluation”, Proceedings of IEEE conference on Cybernetics and Intelligent Systems, 2008, pp. 21-24.
[3] A. C. Boucouvalas, “Real time text-to-emotion engine for expressive internet communication”, Emerging Communication: Studies on New Technologies and Practices in Communication, G. Riva et al. Eds. IOS Press, 2002, pp. 305-318.
[4] A. C. Boucouvalas and X. Zhe, “Text-to-emotion engine for real time internet communication”, Proceedings of International Symposium on Communication Systems, Networks,and Digital Signal Processing, 2002, pp. 164-168.
[5] Z.-J. Chuang and C.-H. Wu, “Emotion recognition from textual input using an emotional semantic network”, Proceedings of International Symposium on Spoken Language Processing, 2002, pp. 177-180.
[6] J. F. Cohn and G. S. Katz, “Bimodal expression of emotion by face and voice”, Proceedings of the 6th ACM International Multimedia Conference Workshop on Face/Gesture Recognition and Their Applications, Sep, 1998, pp. 41-44.
[7] R. Cowie, E. Douglas-Cowie, N. Tsapatsoulis, G. Votsis, S. Kollias, W. Fellenz, and J. Taylor, “Emotion recognition in human-computer interaction”, IEEE Signal Processing Magazine, vol. 18(1), 2001, pp. 32-80.
[8] L. C. De Silva and P. C. Ng, “Bimodal emotion recognition”, Proceedings of 4th IEEE International Conference on Automatic Face and Gesture Recognition, IEEE Computer Society, 2000, pp. 332-335.
[9] L. Devillers, I. Vasilescu, and L. Lamel, “Annotation and detection of emotion in a task-oriented human-human dialog corpus”, Proceedings of ISLE Workshop on Dialogue Tagging for Multi-Modal Human-Computer Interaction, 2002.
[10] L. Devillers, L. Lamel, and I. Vasilescu, “Emotion detection in task-oriented spoken dialogues”, Proceedings of International Conference on Multimedia and Expo, vol. 3, IEEE Computer Society, 2003, pp. 549-552.
[11] K. Forbes-Riley and D. J. Litman, “Predicting emotion in spoken dialogue from multiple knowledge sources”, Proceedings of Human Language Technology Conference of the North American Chapter of the Association for Computational Linguistics, 2004, pp. 201-208.
[12] J. T. Larsen, A. P. McGraw, J. T. Cacioppo, “Can People Feel Happy and Sad at the Same Time”, Journal of Personality and Social Psychology, Vol. 81, 2001, pp. 684-696.
[13] M. Lee, S. S. Narayanan, and R. Pieraccini, “Combining Acoustic and Language Information for Emotion Recognition”, Proceedings of 7th International Conference on Spoken Language Processing, 2002, pp. 873-876.
[14] C. Maaoui, A. Pruski, F.Abdat, “Emotion recognition for human-machine communication”, Proceedings of IEEE/RSJ International Conference on Intelligent Robots and Systems, 2008, pp. 1210-1215.
[15] T. Nguyen, and I. Bass, “Investigation of Combining SVM and Decision Tree for Emotion Classification”, Proceedings of 7th IEEE international Symposium on Multimedia, 2005, pp. 540-544.
[16] Ortony, G. Clore, and A. Collins, The Cognitive Structure of Emotions, Cambridge Press, 1988.
[17] R. W. Picard, Affective Computing. Cambridge, MA: MIT Press, 1997.
[18] R. W. Picard, E. Vyzas, and J. Healey, “Toward machine emotional intelligence: analysis of affective physiological state”, IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 23(10), 2001, pp. 1175-1191.
[19] P. Subasic and A. Huettner, “Affect Analysis of Text Using Fussy Semantic Typing”, IEEE Transactions on Fussy System, vol. 9(4), 2001, pp. 483-496.
[20] T. Wilson, J. Wiebe, and R. Hwa, “Just how mad are you? Finding strong and weak opinion clauses”, Proceedings of 21st Conference of the American Association for Artificial Intelligence (AAAI 04), 2004, pp 761-769.
[21] C.-H. Wu, Z.-J. Chuang, and Y.-C. Lin, “Emotion Recognition from Text Using Semantic Labels and Separable Mixture Models”, ACM Transactions on Asian Language Information Processing, vol. 5(2), 2006, pp 165-183.
[22] T. Yanaru, “An emotion processing system based on fuzzy inference and subjective observations”, Proceedings of 2nd New Zealand Two-Stream international Conference on Artificial Neural Networks and Expert Systems, 1995, pp. 15-20.
[23] Yang, K. H.-Y. Lin, and H.-H. Chen, “Emotion classification using web blog corpora”, Proceedings of IEEE/WIC/ACM International Conference on Web Intelligence, 2007, pp. 275-278.
[24] J. Yang, D. B. Bracewell, F. Ren, and S. Kuroiwa, “The Creation of a Chinese Emotion Ontology Based on HowNet”, Engineering Letters, vol. 16(1), 2008, pp. 166-171.
[25] J. Yang, D. B. Bracewell, F. Ren, and S. Kuroiwa, “A semantic analyzer for aiding emotion recognition in Chinese”, Lecture Notes of Artificial Intelligence, vol. 4114, 2006, pp. 893-901.
[26] Z. Teng, F. Ren, and S. Kuroiwa, “Recognition of Emotion with SVMs”, Lecture Notes of Artificial Intelligence, vol. 4114, 2006, pp. 701-710.
[27] J. M. You, and K. J. Chen, “Automatic Semantic Role Assignment for a Tree Structure”, Proceedings of 3rd SIGHAN Workshop on Chinese Language Processing, 2004, pp. 109-115.
[28] E. T. Robert, The Biopsychology of Mood and Arousal, Oxford University Press,1989.
[29] C. Cortes and V. Vapnik, “Support-vector network”, Machine Learning, vol. 20, 1995, pp. 273-297.
[30] C.D. Manning and H. Schütze, Foundations of Statistical Natural Language Processing, Cambridge, Massachusetts: MIT Press, 1999.