Exploring characteristics of online news comments and commenters with machine learning approaches

摘要

Despite the pervasiveness and significant influence of online news comments, few studies have examined people's commenting behaviors regarding online news articles. This study examined the distribution of age and gender of online news commenters according to news topics. Using a computational approach, we collected publicly available data on online news comments and commenters from the most popular South Korean news aggregator during a three-month period spanning May to July 2016. The clusters of news were identified and categorized using machine learning techniques. The comments and commenters of 20,929 news articles were examined. We found that there were age and gender differences in commenting behaviors that varied based on news topics. Findings concerning large discrepancies in the ages and genders of commenters suggest that online commenting systems should be improved in a way that can guarantee more diverse opinions from readers.