Copy the page URI to the clipboard
Dietze, Stefan and Domingue, John
(2009).
URL: http://semsensweb.di.uoa.gr/semsensweb2009_submiss...
Abstract
The increasing availability of sensor data through a variety of sensor-driven devices raises the need to exploit the data observed by sensors with the help of formally specified knowledge representations, such as the ones provided by the Semantic Web. In order to facilitate such a Semantic Sensor Web, the challenge is to bridge between symbolic knowledge representations and the measured data collected by sensors. In particular, one needs to map a given set of arbitrary sensor data to a particular set of symbolic knowledge representations, e.g. ontology instances. This task is particularly challenging due to the potential infinite variety of possible sensor measurements. Conceptual Spaces (CS) provide a means to represent knowledge in geometrical vector spaces in order to enable computation of similarities between knowledge entities by means of distance metrics. We propose an ontology for CS which allows to refine symbolic concepts as CS and to ground instances to so-called prototypical members described by vectors. By computing similarities in terms of spatial distances between a given set of sensor measurements and a finite set of prototypical members, the most similar instance can be identified. In that, we provide a means to bridge between the real-world as observed by sensors and symbolic representations. We also propose an initial implementation utilizing our approach for measurement-based Semantic Web Service discovery.