Fig 1.
Research framework.
Fig 2.
Trend of comment quantity fluctuation over time.
Fig 3.
The variation in bullet subtitle quantity as it relates to the progress bar.
Table 1.
Topic-word distribution of comment data from the first five days.
Fig 4.
The variation in coherence scores relative to the number of topics in the comment data from the first five days.
Fig 5.
The distribution of comment quantities across different topics based on the first five days of comment data.
Table 2.
Topic-word distribution of comment data from the first month.
Fig 6.
The variation in coherence scores with the number of topics in the comment data over the first month.
Fig 7.
The distribution of comment quantities across different topics based on the first month’s comment data.
Fig 8.
The variation in coherence scores with the number of topics in the bullet subtitle data.
Table 3.
Topic-word distribution of the bullet subtitles data.
Fig 9.
The number of bullet subtitles across different topics.
Table 4.
Prompt design.
Table 5.
Definitions of emotion categories.
Table 6.
Emotion distribution over the first five days.
Fig 10.
The positive and negative percentages for the six topics of comments from the past five days.
Fig 11.
The number of positive and negative comments over the first five days.
Fig 12.
The proportion of positive and negative sentiments in comments from the first five days.
Fig 13.
Distribution of emotional categories in bullet comments.
Fig 14.
The proportion of positive and negative sentiments across the three topics of bullet subtitles.
Fig 15.
The average sentiment of bullet subtitles varies with the progress bar (Note: In the interval [220, 230), no bullet subtitles are present, resulting in a sentiment value of 0).
Fig 16.
The cumulative sentiment of bullet subtitles varies along the progress bar.
Fig 17.
The manner in which the second uncle is ‘gazed at’ by audiences in the short video.