Analysis of Research Trends on ‘Social Isolation’ through Topic Modeling Methodology

This study analyzed domestic research trends on social isolation using topic modeling. Reviewing the previous studies, it examined how reseach trends in social isolation have changed domestically and internationally. For this purpose, the study examined major research topics in domestic academic papers listed in the Korea Citation Index (KCI) database from 2004 to 2024 using social isolation as a keyword. As a research method, we utilized Latent Dirichlet Allocation (LDA), which is an effective method for analyzing unstructured text data to derive major topics appearing in a document set. As a result, first, most domestic social isolation studies deal with loneliness or social isolation of vulnerable groups such as the elderly, children, adolescents, women, immigrants, and the disabled. In particular, topics related to the elderly account for an overwhelmingly high proportion. Second, it was found that most studies have focused on social isolation in relation to lonely death, loneliness, and depression, indicationg a prominent tendency to perceive social isolation at the individual level rather than the societal level. Third, research topics related to social structure or political system were found to be so weak that they were not classified as topics. Fourth, we found that most of the research is focused on case studies related to individual mental health, communities, which means the in-depth reseaches from the social structural perspective focusing on social inequality, urbanization, or globalization have not been sufficiently conducted. Social isolation is emerging as a significant social problem that affects the entire society beyond the subjective emotions of individuals such as loneliness. This study aims to provide implications for future policy development and academic research directions by examining domestic research topics on social isolation.