Expression and discussion: Sentiment analysis based on the text of forum replies in the case of ‘Songling Lin’

HE Jingjie

Chinese Journal of Journalism & Communication ›› 2015, Vol. 37 ›› Issue (9) : 109-132.

PDF(1667 KB)
PDF(1667 KB)
Chinese Journal of Journalism & Communication ›› 2015, Vol. 37 ›› Issue (9) : 109-132.

Expression and discussion: Sentiment analysis based on the text of forum replies in the case of ‘Songling Lin’

  • He Jingjie is a teacher at the School of Presenting Art, Communication University of China. Email: hejingjie0823@163.com.
Author information +
History +

Abstract

Based on the heatedly discussed social incident - “six policemen in Harbin beat a young man to death”, this study concentrates on analyzing the text of relevant comments posted by Chinese netizens on the forum of “Mop”. In order to figure out how the Chinese netizens behave online when discussing prominent and controversial social issues, the author has utilized the method of sentiment analysis and divided the research object into two dimensions - the netizens’ individual expression and the responded-to discussion among them. By respectively examining the length, punctuation and opinion-topic relation of the forum replies and measuring their sentiment polarities, the results revealed that most netizens’ self-oriented replies tend to express negative, radical and extreme emotions toward the government and authorities with only a few words. During the process of online discussion, the relationship among participants is more likely to grow into antagonism. Moreover, the netizens involved in debates tend to behave in a more aggressive manner, thus most debates eventually deviate from the principal contradiction and turn into personal attacks.

Key words

Sentiment analysis / text of forum replies / netizens’ online behavior

Cite this article

Download Citations
HE Jingjie. Expression and discussion: Sentiment analysis based on the text of forum replies in the case of ‘Songling Lin’[J]. Chinese Journal of Journalism & Communication. 2015, 37(9): 109-132
PDF(1667 KB)

Accesses

Citation

Detail

Sections
Recommended

/