Topic Modeling of Indonesian Children's Literature Using Latent Semantic Analysis

The Covid-19 pandemic resulted a significant reduction in the number of publications of literary works, especially works of children's literature, which once dominated the first ranking of the most popular subjects at the Indonesian Publishers Association. So that efforts are needed to raise the enthusiasm of Indonesian children's literature. In addition, data on children's literature is scattered across various portals that compile metadata/bibliography and synopsis summaries of children's literature stories from various sources or online portals. A portal that collects data on all Indonesian children's literature that is grouped automatically according to the most dominant topics reflecting the context of each literary work is urgently needed to facilitate search, develop teaching materials, and inspire the creation of new literary works. This study uses qualitative methods and experiments with several stages of research including metadata analysis of models of children's literature, secondary data collection, topic modeling of children's literature with Latent Semantic Analysis (LSA). The results of the study showed that the topic clustering was most often raised by children's authors published on the Mizan.