Towards a Unified Ontology for Software Defect Classification in African Financial Technology Systems

The absence of a unified defect classification ontology tailored to the financial technology domain in Africa creates significant barriers to cross-organisational defect learning, regulatory reporting, and comparative research. This dissertation develops and validates a unified software defect classification ontology for African fintech systems, with primary grounding in the Nigerian fintech ecosystem of Lagos, South West Nigeria. A design science research methodology is employed across four artefact design iterations, combining ontology engineering techniques with empirical validation. Phase one conducts a systematic literature review of 95 existing defect taxonomies and classification standards including IEEE 1044, ODC, and Hewlett-Packard defect categorisation. Phase two analyses 6,500 production defect reports from five Nigerian fintech companies using open and axial coding to identify domain-specific defect categories absent from existing ontologies. Phase three constructs the Nigerian and African Fintech Defect Ontology (NAFDO) using OWL 2 and validates it through ontology evaluation metrics including competency question coverage and reasoner consistency. Phase four deploys NAFDO as a classification tool at two fintech firms over six months, measuring inter-rater classification agreement and cross-firm defect pattern analysis capability. Results show NAFDO achieves 0.81 Cohen's kappa for inter-rater agreement and enables novel cross-firm defect benchmarking not possible with existing taxonomies. The dissertation provides the first fintech-domain defect ontology developed from African production data. Keywords: defect classification, ontology, fintech, Nigeria, software engineering theory