<![CDATA[In 2023, the National Library of Indonesia, through its Library Analysis and Reading Culture Development Centre, organised several literacy talk shows. Each event was documented in minutes (.doc or .pdf format) recording the speakers' material, questions, and answers. As events increased, so did the volume of minutes. This research aimed to identify frequently discussed topics using Text Mining with a Topic Modelling approach. Latent Dirichlet Allocation was applied and evaluated by perplexity values (a measure of model quality). Results showed the optimal number of topics to represent the dataset was three, with the lowest perplexity value of 470.922 at the 30th iteration. The three main topics identified were reading interest and the need for books in schools and regions, libraries’ role in improving children’s literacy, and librarians' role in inclusive literacy programmes for both young and old, including health literacy. Frequent words were literacy, library, reading, books, and children.]]>
Copyrights © 2024
