Development of a domain ontology to support information retrieval on the South African informal sector services

Abstract

The high unemployment rate in South Africa has encouraged several types of informal occupational and business activities in the informal business sector. Like in many Sub-Saharan Africa (SSA) countries, the South African informal business sector (SAIBUS) has many challenges to contend with, such as a lack of organisation, poor regulation, poor service delivery, low capitalisation, and low productivity. Attempts have been made to address some of these challenges by applying ICTs. One proposed solution is using web portal technology to promote informal businesses and facilitate informal business transactions. However, to attain maximum impact, the platform must be enabled with Artificial Intelligence (AI) to create a system that can leverage the convergence of intelligent reasoning and semantic technology in tackling the challenges of the South African informal business sector. In essence, such a web portal must possess capabilities such as semantic-based reasoning, semantic analysis of user queries, and intelligent recommendations, which require the existence of a dedicated domain ontology. Therefore, this study aimed to develop a domain ontology that can support information retrieval and promote intelligent reasoning on services in the informal business sector in South Africa. Based on the study's aim, the objectives were: (1) to determine the requirements for developing a domain ontology for the South African informal business sector; (2) To design and develop a domain ontology for the South African informal business sector; (3) To evaluate the quality attributes of a domain ontology for the South African informal business. Design Science Research (DSR) was employed as a research strategy to accomplish the objectives of this research. The research design mapped the phases of DSR to the Lean Ontology Development (LOD) method to realise the phases of problem awareness, design, development, and evaluation. While reviewing some use cases from the South African informal business sector, we used competency questions to identify the requirements for the South African informal business sector (SAIBUS). The LOD method was adopted as the ontology engineering methodology. A combination of ontology modularisation and design patterns were used as concepts to realise the design of the SAIBUS ontology, while the Protégé ontology editor was used to develop the ontology. The evaluation of the ontology was divided into two phases – validation and verification. The validation of the SAIBUS ontology (assessing its quality and correctness) was done using the web-based tool - OOPS! and the HermiT ontology reasoner. The SAIBUS ontology was verified by comparing it against the ontology requirements and competency questions to assess whether the ontology is built to fit its purpose. This was done by attempting to answer some of the competency questions using the SPARQL query to test its ability to retrieve relevant ontological instances (individuals). The results from the evaluation show that the SAIBUS ontology can support information retrieval on the South African informal sector services.