CC BY 4.0Salcedo, Miriam Charmigrace Q.Bagunu, Jessie Rose M.Amandy, Michael D.Tinao, Maria Maura S.2026-09-012026-09-012026-09-01https://2026.ifla.orghttps://repository.ifla.org/handle/20.500.14598/7325This study explores how subject access practices in academic libraries shape the visibility and discoverability of LGBTQIA+ knowledge, focusing on systemic gaps in representation within the Philippine academic landscape. Standardized subject headings and legacy classification schemes often fail to capture the nuanced complexity and evolving nature of LGBTQIA+ identities. Analyzing over 1,400 bibliographic records across five library units at the University of the Philippines Diliman, this study employs a mixed-methods research design that combines computational natural language processing with qualitative insights from librarian interviews and user feedback. The methodology implements a three-stage computational pipeline: keyword extraction using RAKE and TF-IDF to generate high-impact descriptors, thematic exploration via K-Means clustering to reveal latent structures, and semantic retrieval using Cosine Similarity to build relationship-based pathways. Findings reveal that while relevant resources are present, discoverability is severely constrained by generalized terminology, missing metadata (with ~73% of records lacking summaries), severe keyword flattening, and record duplication. To address these structural barriers, the paper proposes a responsive subject analysis framework that integrates localized vocabularies, multi-dimensional keywording practices, and AI-assisted metadata enrichment to establish equitable knowledge systems.enghttps://creativecommons.org/licenses/by/4.0/LGBTQ library servicesDiscoverabilityArtificial intelligenceAcademic and research librariesSubject accessMetadataEvents materialtMiriam Charmigrace Q. SalcedoJessie Rose M. BagunuMichael D. AmandyMaria Maura S. Tinao