Deliverable 4.2 – Model annotation components and guidelines

Posted by patrick_maue on January 30, 2011 under Deliverable | Be the First to Comment

Link to the PDF file:

D4.2 – Model annotation components and guidelines

Executive Summary:

This document describes the methodology we propose to semantically annotate environmental models, as well as the components that have been developed to perform the annotation process. Domain ontologies are one of the key elements in this process, hence we added at the end of the deliverable a discussion about how to induce domain ontologies from text and streams. In order to apply the guidelines described here to a real use case, we created the ENVISION simple use case (see section 1). It consists of a service composition including different OGC Web services: a Sensor Observation Service (SOS), a Web Feature Service (WFS) and a Web Processing Service (WPS).

In the  second  section of the document we  extend the semantic annotation concepts we introduced in Deliverable 4.1 by adding new services like the Web Coverage Service (WCS), and by giving more details about annotations for environmental models. For each described service, a generic approach is presented considering the various representations taking part in the process: the Capabilities document, the Procedure-Oriented Service Model (POSM),  the Web Service Description Language (WSDL), and  the application ontology representing the data model (RDF data model). The annotation  examples created for the ENVISION simple use case are also described to close the section. Different software components are required to proceed with the annotations within the ENVISION infrastructure. This paper presents some new developments (e.g. Service Model Translator (SMT)), and tools based on previous implementations from the SWING project (e.g. Web OntoBridge).

In section 3, we explain how these components are integrated in the architecture and how to use them. In the fourth section we present ontology induction from text and streams along with the results of using it to induce domain ontologies. We introduce  the OntoGen Tool which is used for semi- automatic ontology induction, and illustrate its applicability on the data relevant for Envision. We have shown potential usage of semi-automatic ontology creation from the domain specific data for the general Envision domain and for one of the Envision case studies.