This study aims to empirically analyze the structure and temporal
evolution of population discourse in Korean media by applying text mining
techniques to 574,189 news articles published between 1990 and 2024.
Utilizing multiple statistical language mo...
This study aims to empirically analyze the structure and temporal
evolution of population discourse in Korean media by applying text mining
techniques to 574,189 news articles published between 1990 and 2024.
Utilizing multiple statistical language models—including Term Frequency–
Inverse Document Frequency (TF-IDF), Latent Dirichlet Allocation (LDA), and
static topic modeling—this study seeks to identify how key population-related
issues have been formed and transformed within the social context, and how
population discourse has shifted in its macro-level orientation over time.
The findings provide three main contributions. First, the study expands the
analytical scope of population issues beyond the narrow focus on declining
fertility or population aging, interpreting them within the broader
structural transformations of Korean society. Second, by integrating TF-IDF,
LDA, and static topic modeling, it proposes a multilayered analytical
framework that captures three dimensions of discourse evolution—yearly
keywords, thematic structures, and inter-topic relationships. Third, through
a quantitative examination of inter-topic correlations over a long-term
timespan, the study empirically demonstrates how population discourse has
been shaped and restructured through interactions among diverse policy
domains.