ABSTRACT
We examined how violent scenes and words used in action- and non-action-oriented webcomics were associated with the frequency by which violent words were used in readers’ online comments. Data on the webcomics provided by Naver, a popular webcomics provider, were analyzed with computational approaches. Primary results were as follows: there was a positive relationship between the number of violent words used in a webcomic and that of violent comments regardless of genre, and the effects of violent scenes on the number of violent comments were statistically insignificant regardless of the types of webcomics.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Notes
2 The accuracies of VGG19, MobileNet, ResNet152, and DenseNet201 were 87.38, 85.44, 84.47, and 83.50, respectively.
3 The sites providing such a dictionary are https://asus1004.tistory.com/241, https://github.com/organization/Gentleman/blob/master/resources/badwords.json, and https://github.com/Hanul/BadWordFilter/blob/master/DB/KO.
5 These analyses were carried out using the final dataset composed of webtoon episodes without outliers with respect to the number of violent words used in readers’ comments.
Additional information
Funding
Notes on contributors
Sang Yup Lee
Sang Yup Lee (Ph.D, Michigan State University) is an Associate Professor at Communication Department, Yonsei University. His research interests include media psychology, computational social science, and data science.
Min Yeob Kim
Min Yeob Kim is a graduate student at Digital Analytics, Yonsei University. His research interests include media industry, computer vision, deep learning, and data science.
Pil Kyu Choi
Pil Kyu Choi is a graduate student at Digital Analytics, Yonsei University. His research interests include computer vision, recommendation systems, and data science.