Algorithmic Approaches to Subject Indexing of Nigerian Language Materials in Multilingual Library Databases

This study investigated algorithmic approaches to subject indexing of materials in Nigerian languages, specifically Yoruba, Hausa, Igbo, and Fulfulde, within multilingual library databases. Existing library databases are designed predominantly around English-language materials, creating systematic indexing gaps that render indigenous language materials invisible or poorly retrievable in catalogue searches. The study adopted a design science research paradigm, combining computational linguistics, natural language processing, and library indexing theory. Prototype subject indexing algorithms were developed for each of the four target languages and tested against a corpus of 2,000 items in the test collection of the National Library of Nigeria. Retrieval precision and recall were compared against manual indexing benchmarks. The study found that language-specific morphological analysis significantly improved indexing accuracy for all four languages. A generalised multilingual indexing framework for Nigerian language collections was developed and validated. The study makes original contributions to multilingual information retrieval, computational cataloguing, and the preservation of linguistic diversity in national library systems. Keywords: subject indexing, Nigerian languages, multilingual databases, natural language processing, library cataloguing.