Empathi: An Ontology for Emergency Managing and Planning About Hazard Crisis

Abstract

In the domain of emergency management during hazard crises, having sufficient situational awareness information is critical. It requires capturing and integrating information from sources such as satellite images, local sensors and social media content generated by local people. A bold obstacle to capturing, representing and integrating such heterogeneous and diverse information is lack of a proper ontology which properly conceptualizes this domain, aggregates and unifies datasets. Thus, in this paper, we introduce empathi ontology which conceptualizes the core concepts concerning with the domain of emergency managing and planning of hazard crises. Although empathi has a coarse-grained view, it considers the necessary concepts and relations being essential in this domain. This ontology is available at https://w3id.org/empathi/.

A Pattern Driven Approach to Knowledge Representation in the Disaster Domain

“…To support the interdependent roles amongst actors, multiple international agencies have long emphasized the importance of a unified understanding of the disaster management domain by creating common glossaries, vocabularies and thesauri [10,11]. However, these are not expressive enough to explicitly capture domain knowledge [5]. This makes it difficult to retrieve and integrate data/information meaningfully.…”

“…In the Hazard, Vulnerability Risk (HVR) portion of the disaster management cycle, work has been done to conceptualize hazard [12], vulnerability [13] and risk [14][15][16] concepts. Additionally, surveys conducted by [5,6] show that there exists several formal ontologies and vocabularies for capturing, representing and integrating heterogeneous data in the emergency management sub-domain. Artificial Intelligence (AI) logic rules lack expressivity to sufficiently capture domain knowledge [17].…”

“…Typically, the creation of new ontologies is either motivated by the fact that existing ontologies and terminologies are discipline-specific and consequently not comprehensive [7,8] or by the lack of expressive power to capture requirements in the disaster management domain [5,23]. Therefore, the commitments and possibly complex alignments associated with these disaster ontologies make them not easily reusable in other contexts.…”

“…The EventClassification pattern uses DOLCE foundational ontology concepts with a minimal number of classes and axioms to avoid over-commitment and support pattern Fig. 3 Instantiating of ontology design patterns to organize meteorological data 5 Find EventClassification pattern implementation at https ://w3id. org/gicen tre/onto/Event Class ifica tion.owl.…”

“…The top-down ontology modelling approach based on foundational ontologies has been widely used as a common denominator in the disaster domain [5,6]. Such ontologies are associated with undesired commitments that will often make their reuse in other contexts difficult.…”