Published September 27, 2011 | Version v1
Project deliverable Open

Characterisation mechanisms for unknown data sources. PlanetData Deliverable D1.1

Authors/Creators

  • 1. ROR icon Universidad Politécnica de Madrid
  • 2. ROR icon Jožef Stefan Institute
  • 3. ROR icon École Polytechnique Fédérale de Lausanne

Description

We describe a set of characterisation mechanisms focused on two data modalities: sensor-based and text based stream data sources. By characterisation we refer to the process of semantically annotating data sources according to some domain or general ontologies, and/or linking these data sources to external publicly avaiable knowledge sources (e.g. DBpedia). In the area of sensor data streams, we present work done for the ontology-based characterisation of environmental sensors from the SwissExperiment deployment in Switzerland, and for the characterisation of streaming data sources from the Pachube platform. We also describe the incipient work on the so-called Linked Stream Data or Linked Sensor Data, and present two alternative approaches to generate it. In the area of text stream characterisation, we describe techniques used for the characterisation of Web2.0 data sources, such as microblogs, by means of topic recognition and clustering. 

Files

D1.1_Characterisation_mechanics_for_unknown_data_sources.pdf

Files (3.4 MB)

Additional details

Funding

European Commission
PLANETDATA - PlanetData 257641