📖 ABSTRACT/OVERVIEW
The Nigerian legal corpus, comprising case law, statutes, regulations, and legal commentary, is voluminous, poorly digitised, and largely inaccessible to machine processing, limiting the development of legal AI tools for Nigerian practitioners and researchers. This study develops an original knowledge graph construction and reasoning methodology for Nigerian legal document analysis. Three original contributions are presented: a Nigerian legal ontology (NigeriaLegalOnt) formalising entity types (cases, statutes, courts, judges, legal concepts, parties) and semantic relationships specific to Nigerian common law and legislation; a hybrid information extraction pipeline combining rule-based pattern matching for structured legal fields with a fine-tuned LegalBERT-NG model (trained on 240,000 Nigerian court judgments) for entity and relation extraction from unstructured judicial text; and a neurosymbolic reasoning module integrating the knowledge graph with a legal case similarity reasoning algorithm based on graph embedding and analogical case retrieval. The knowledge graph was populated with 18,500 Supreme Court and Court of Appeal judgments and 3,200 statutory instruments. Entity extraction F1 reached 0.86 for case citations and 0.79 for legal concept identification. Case similarity retrieval achieved a mean average precision of 0.73 in expert evaluation by ten legal researchers. The study constitutes an original contribution to legal AI, legal knowledge representation, and the nascent field of computational Nigerian legal scholarship.
Keywords: knowledge graph, Nigerian legal AI, information extraction, legal ontology, legal reasoning
Need Complete Chapters of the Above Topic?
Get high-quality, Zero-AI research materials with current citations.
Request via WhatsApp 💬