UniNMS: Design, Implementation, and Evaluation of a Hybrid Semantic-Lexical Knowledge Management System for Nigeria's National University Research Network

Nigeria’s 157 NUC-accredited universities produce over 18,000 research papers annually yet lack any unified repository, cross-institutional search capability, or domestic journal hosting infrastructure. This fragmentation results in an estimated ₦15–20 billion in annual duplicated research expenditure and a national citation impact of 1.8 citations per paper, 65% below the global average of 5.2. This paper presents UniNMS (University National Management System), a full-stack knowledge management system designed for Nigeria’s institutional, linguistic, and infrastructure context. The system comprises a Next.js 14 web application (52 pages, six role contexts), an Expo SDK 54 mobile application (13 screens), and an Express.js REST API (131 endpoints across 22 route files) backed by PostgreSQL 16 with the pgvector extension. Three principal contributions are evaluated: (i) a three-pass hybrid plagiarism detection pipeline achieving precision 0.92, recall 0.95, and F1 0.94; (ii) a hybrid semantic-lexical retrieval system with empirically optimised weighting (α = 0.60) yielding F1 0.84 on a 363-document Nigerian corpus; and (iii) a module implementing Free, Prior and Informed Consent (FPIC) and the CARE Principles (Collective Benefit, Authority to Control, Responsibility, and Ethics) for indigenous data governance. Administrator accounts have been provisioned for all 157 NUC-accredited institutions; the platform includes Unstructured Supplementary Service Data (USSD) gateway access (*123#) for low-bandwidth campuses. Findings demonstrate the feasibility of sovereign, nationally-appropriate academic knowledge infrastructure for Nigerian higher education.