The Role of a Multitier Ontological Framework in Reasoning to Discover Meaningful Patterns of Sustainable Mobility

The successful applications of Geographical Knowledge Discovery in Databases (GKDD) are not common, despite the vast literature on knowledge discovery in databases. The reason is that, although it is relatively straightforward to find patterns in very large spatio-temporal databases, establishing their relevance and explaining their causes are both very complex problems. In practice, spatio-temporal databases are not adequate to adopt geographical knowledge in an ad-hoc manner, and as a result, most of the patterns found in a GKDD process may well already be background knowledge of an application domain. Addressing these issues requires considering a knowledge discovery process as a multi-tier ontological process, in the sense that more complex reasoning modes need to be used to help the comprehension of what makes one pattern structurally and meaningfully different from another. Towards this goal, this chapter proposes a multi-tier ontological framework to support a GKDD process. Three ontological tiers are described to provide the common base for the organisation of different nature and sources of knowledge as well as the reasoning tasks integrated within a spatio-temporal database. The potential of this approach is illustrated on (1) reducing the semantic gap between ontological and database representation by using mappings between ontology and conceptual model, and (2) permitting to define spatio-temporal relationships within an ontology, which can be translated to spatio-temporal conceptual queries. The results are drawn from one specific metaphor: the movement-as-trajectory. Our preliminary results are pointing out that knowledge of sustainable mobility must come from a global and systemic view of patterns within a GKDD process, and the important role of a multi-tier ontological framework on the integration of semantic abstractions (concepts), reasoning tasks, and patterns. They had also drawn attention to the fact that combining ontological representation with database querying mechanisms is fundamental for the use of ontologies in GKDD processes