Matthias Moi, Nikolai Rodehutskors
This document deals with the general approach of the EmerGent ontology and the software component SemanticDataStore. The EmerGent ontology is needed to model and process emergency related information semantically. The document provides an overview of existing standards and frameworks for semantic data processing. Furthermore, existing information models like vocabularies, taxonomies and ontologies from different perspectives are analysed. The analysis includes domain independent models, e.g. to model people or general resources, as well as domain dependent models, e.g. to model dangers or incidents and resources. Furthermore, this document deals with the EmerGent ontology itself. It describes the requirements placed in the ontology, the structure of the ontology and explains, why certain existing standards and ontologies were used and implemented in the EmerGent ontology. The EmerGent Alerts and especially their structure and the research, which led to these alerts, are explained. Further introduced parts of the ontology, such as the information quality graphs, which are important for the quality evaluation of user generated contend are described. From the technical perspective this deliverable concerns software frameworks to develop the SemanticDataStore and technologies to achieve scalability in order to manage massive data gathered from social media. The API used in the SemanticDataStore is described and possible ways for the usage of the API are explained. Regarding the requirements placed in the SemanticDataStore, it has been benchmarked using real data in order to assess its scalability.
Purpose of the Document
This deliverable contributes to objective O2 – Show the positive impact of information mining, information quality, information gathering and information routing for social media in emergencies. The main objective of this deliverable is to build the basis for the development of the ontology and the software component SemanticDataStore in order to model and store emergency related information from social media.