Propose: In the era of digital transformation, research in the field of public libraries—recognized as a foundation for cultural, social, and educational development—requires systematic and predictive analysis to identify emerging trends and future directions. This study aims to analyze thematic trends and forecast the future of public library research through an integrated text mining and deep learning approach.
Method: The dataset comprised more than 7,000 scholarly articles indexed in Scopus and the Islamic World Science Citation Center (ISC) from 1889 to 2024. Latent topics were identified using the Latent Dirichlet Allocation (LDA) algorithm and categorized into eight thematic clusters at the international level and five clusters at the national level. Subsequently, the publication trends of each cluster were modeled using a Long Short-Term Memory (LSTM) recurrent neural network to predict developments up to 2034 for international research and 1413 (2034–2035) for national research.
Findings: The findings indicated substantial growth in themes such as “Digital Technologies and Intelligent Information Services” and “Health Literacy and Social Service Enhancement,” whereas more traditional areas, including “Professional Challenges of Librarians,” exhibited a declining trend. These results highlight a growing international shift in public library research toward data-driven and AI-oriented services.
Originality/Value: The results indicate that the evolution of research in the field of public libraries aligns with the global shift toward data-driven services and artificial intelligence. This study provides a model for scientific foresight and data-informed decision-making within the national public library system.