使用者經由發佈社群網路文章，來紓發日常的心得與觀察，已是再平常不過的事。藉由文章當中語意模式的截取方法，可以找出文章裡所表露的情緒，在本研究中進一步探索隱藏性情緒語意模式(亦即不帶有字典中定義的顯性情緒字的語意模式)在不同的情緒上的具有何特質，其情緒分數與模式之間的關聯如何。透過語意模式清單的建置，針對快樂、生氣、憂傷、恐懼、希望以及驚喜這六種情緒，探索在各別的情緒上，這些不帶有明顯情緒字的語意模式，具體是如何，而當中又有多少與特定的情緒有高度相關。研究結果顯示不帶有情緒字的隱藏性情緒語意模式可以幫助判別訊息中的單一情緒。

With the widespread usage of social media, the amount data that these platforms generate has made it possible to analyze the emotions and opinions within text messages. Most of the existing research related to emotion classification utilize a set of agreed-upon emotion specific words, and ignore the presence of stop words; which were supposed to bear no emotional meaning when used on their own. In this research we present our work on Implicit Emotion Patterns, which are composed only of stop words. The extracted patterns are highly relevant to a specific emotion in the set of joy, anger, sadness, surprise, hope, or fear; based on their emotion score. We set thresholds on the emotion score for each emotion to filter out these patterns. To evaluate, we prepare an experiment with a sample of tweets containing these Implicit Emotion Patterns for each of the emotions, and examine their relationship with the emotions outline. We asked annotators to check the agreement of the emotions they perceive in the tweets. We find that these Implicit Emotion Patterns are highly related to the emotions they were relevant to. In this work, we contribute to the domain a dictionary of stop words that display a relationship with emotions and outline our method for examining this hypothesis.

摘要

i

Abstract

i

Acknowledgement

ii

List of Tables

v

List of Figures

vi

1 Introduction

1

2 Related Work

3

3 Methodology

8

3.1

Framework . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

3.2

Emotion Word . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

3.3

Pattern Extraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

3.4

emotion specific pattern Selection . . . . . . . . . . . . . . . . . . . . . . 13

3.5

Data Collecting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

3.6

Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

4 Experiment & Results

16

iii

5 Discussion

26

6 Conclusion

30

References

31

[1] Apoorv Agarwal, Boyi Xie, Ilia Vovsha, Owen Rambow, and Rebecca Passonneau.
Sentiment analysis of twitter data. In Proceedings of the workshop on languages in
social media, pages 30–38. Association for Computational Linguistics, 2011.
[2] Carlos Argueta, Fernando H Calderon, and Yi-Shin Chen. Multilingual emotion classifier using unsupervised pattern extraction from microblog data. Intelligent Data
Analysis, 20(6):1477–1502, 2016.
[3] R Sherlock Campbell and James W Pennebaker. The secret life of pronouns flexibility
in writing style and physical health. Psychological science, 14(1):60–65, 2003.
[4] Niladri Chatterjee and Neha Kaushik.

Rent: Regular expression and nlp-based

term extraction scheme for agricultural domain. In Proceedings of the International
Conference on Data Engineering and Communication Technology, pages 511–522.
Springer, 2017.
[5] Dmitry Davidov, Oren Tsur, and Ari Rappoport. Enhanced sentiment learning using
twitter hashtags and smileys. In Proceedings of the 23rd international conference on

31

computational linguistics: posters, pages 241–249. Association for Computational
Linguistics, 2010.
[6] Louis August Gottschalk and Goldine C Gleser. The measurement of psychological
states through the content analysis of verbal behavior. Univ of California Press, 1969.
[7] Francis Guillemin, Claire Bombardier, and Dorcas Beaton. Cross-cultural adaptation
of health-related quality of life measures: literature review and proposed guidelines.
Journal of clinical epidemiology, 46(12):1417–1432, 1993.
[8] Trevor Sheldon Hunte. Term associated emotion classification. , pages 1–27, 2013.
[9] Takayuki Kawabata, Fumiaki Itoh, Haruo Yokota, Yosuke Watanabe, and Qiang Song.
Frequent pattern extraction apparatus frequent pattern extraction method and program, February 16 2016. US Patent 9,262,714.
[10] Efthymios Kouloumpis, Theresa Wilson, and Johanna D Moore. Twitter sentiment
analysis: The good the bad and the omg! Icwsm, 11(538-541):164, 2011.
[11] Chia-Chen Lee and Shu-Kai Hsieh. Evaluative pattern extraction for automated
text generation. In The 9th International Natural Language Generation conference,
page 99, 2016.
[12] Stephen J Lepore and Joshua M Smyth. The writing cure: How expressive writing promotes health and emotional well-being. American Psychological Association,
2002.
[13] Haiying Li, Zhiqiang Cai, Arthur C Graesser, and Ying Duan. A comparative study
on english and chinese word uses with liwc. In FLAIRS Conference, 2012.

[14] Thomas B Mackenzie, Stanley D Rosenberg, Bernard J Bergen, and Gary J Tucker.
The manipulative patient: an interactional approach. Psychiatry, 41(3):264–271,
1978.
[15] Alexander Pak and Patrick Paroubek. Twitter as a corpus for sentiment analysis and
opinion mining. In LREc, volume 10, 2010.
[16] James W Pennebaker. Putting stress into words: Health, linguistic, and therapeutic
implications. Behaviour research and therapy, 31(6):539–548, 1993.
[17] James W Pennebaker, Martha E Francis, and Roger J Booth. Linguistic inquiry and
word count: Liwc 2001. Mahway: Lawrence Erlbaum Associates, 71(2001):2001,
2001.
[18] James W Pennebaker, Tracy J Mayne, and Martha E Francis. Linguistic predictors
of adaptive bereavement. Journal of personality and social psychology, 72(4):863,
1997.
[19] Robert Plutchik. The nature of emotions human emotions have deep evolutionary
roots, a fact that may explain their complexity and provide tools for clinical practice.
American Scientist, 89(4):344–350, 2001.
[20] Samira Shaikh, Laurie Beth Feldman, Eliza Barach, and Yousri Marzouki. Tweet
sentiment analysis with pronoun choice reveals online community dynamics in response to crisis events. In Advances in cross-cultural decision making, pages 345–
356. Springer, 2017.
[21] . A Twitter Corpus for the Evaluation of Keyword Extraction. PhD thesis, , 2012.