Ontology for Agent-Based Modeling of Oral Health Disparities

Abstract

The diversity of dynamic modeling approaches creates challenges for ensuring structural legitimacy of the resulting models. Agent-based modelers have the flexibility to create, analyze, and conduct experiments, but face the problem of structural uncertainty that influences simulation results. The structural uncertainty, related with lack of knowledge of reality, arises from the absence of a general model structure or principle on model design. Ontology provides a structural foundation for agent-based models (ABM) that can improve their correspondence with reality. This dissertation research develops a novel approach to combining ABM and ontology that streamlines model design and implementation. An ABM ontology is developed as the conceptual framework that provides a general model structure with a clear definition of major ABM components and relations between them. This approach is demonstrated by developing comparative models of miasma and germ theories of disease. Having established the utility of the approach, it is used to develop an oral health model to simulate oral health behavior for racial and ethnic minority older adults in an urban environment through related factors and cumulative experiences of individuals in regards to their oral health and dental visits. This modeling effort contributes to inform implementation of effective and appropriate oral health policies and interventions to improve oral health equity by easing the oral disease burden for disadvantaged populations.