Exploiting a perdurantist foundational ontology and graph database for semantic data integration

Abstract

The view of reality that is inherent to perdurantist philosophical ontologies, often termed four dimensional (4D) ontologies, has not been widely adopted within the mainstream of information system design practice. However, as the closed world of enterprise systems is opened to Internet scale Semantic Web and Open Data information sources, there is a need to better understand the semantics of both internal and external data and how they can be integrated. Philosophical foundational ontologies can help establish this understanding and there is, therefore, an emerging need to research how they can be applied to the problem of semantic data integration. Therefore, a prime objective of this research was to develop a framework through which to apply a 4D foundational ontology and a graph database to the problem of semantic data integration, and to assess the effectiveness of the approach. The research employed design science, a methodology which is applicable to undertaking research within information systems as it encompasses methods through which the research can be undertaken and the resultant artefacts evaluated. This methodology has a number of discrete stages: problem awareness; a core design-build-evaluate iterative cycle through which the research is conducted; and a conclusion stage. The design science research was conducted through the development of a number of artefacts, the prime being the 4D-Semantic Extract Load (4D-SETL) framework. The effectiveness of the framework was assessed by applying it to semantically interpret and integrate a number of large scale datasets and to instantiate a prototype graph database warehouse to persist the resultant ontology. A series of technical experiments confirmed that directly reflecting the model patterns of 4D ontology within a prototype data warehouse proved an effective means of both structuring and semantically integrating complex datasets and that the artefacts produced by 4D-SETL could function at scale. Through illustrative scenario, the effectiveness of the approach is described in relation to the ability of the framework to address a number of weaknesses in current approaches. Furthermore the major advantages of the 4D-SETL are elaborated; which include ability of the framework is to combine foundational, domain and instance level ontological models in a single coherent system that dispensed with much of the translation normally undertaken between conceptual, logical and physical data models. Additionally, adopting a perdurantist realist foundational ontology provided a clear means of establishing and maintaining the identity of physical objects as their constituent temporal and spatial parts unfold over the course of time