Analysis of Trends in Train Accident-Related News Using Topic Modeling

Train accidents cause significant human casualties and property damage, and various accidents continue to occur despite technological advancements. The media reports these accidents, contributing to public safety awareness and policy changes. This paper aims to analyze news articles related to train accidents over a decade, from 2014 to 2023, to develop prevention and response strategies. By analyzing the social discourse and media coverage trends of train accidents through news articles, this study comprehensively examines not only the accidents but also changes in societal perception and policy responses. Using LDA topic modeling, we derived five optimal topics: (1) train accidents in South Korea, (2) major international train accidents, (3) screen door accidents, (4) train-related disasters and corresponding measures, and (5) safety measures for preventing train accidents. This research provides a comprehensive perspective on train safety through news article analysis, examining not only the accidents themselves but also changes in social perception and policy responses by analyzing media reporting trends. We expect this study to contribute to the development of train safety policies and guidance for media reporting directions.