Paikkatiedon semanttinen mallinnus, integrointi ja julkaiseminen Case Suomalainen ajallinen paikkaontologia SAPO Tomi Kauppinen, Eero Hyvönen, Jari Väätäinen Semantic Computing Research Group (SeCo) http://www.seco.tkk.fi/ Aalto University Department of Media Technology University of Helsinki Department of Computer Science
Mihin tarvitaan paikkaontologioita? 1. Yleinen kiinnostus esim. navigointiin ja paikkoihin liittyviin palveluihin 2. Paikkojen identifiointi ja tunnistus hankalaa: - paikat voivat olla samannimisiä ja - paikkoja on valtava määrä Semanttisten teknologioiden tuomat mahdollisuudet hakujen tehostajana 30-Mar-10 2
Mihin tarvitaan paikkaontologioita? Resurssit liittyvät tiettyihin paikkoihin, koska esim. - Jokin esine on valmistettu, käytetty tai sen sijainti on Helsingissä - Eräs taulu on maalattu Punkaharjulla ja viimeistelty Porvoossa - Jokin elokuva on kuvattu Helsingissä mutta sen tapahtumat sijoittuvat Pietariin - Eräs valokuva on kuvattu tunturin laella Lapissa -... 30-Mar-10 3
Problems due changes - Content indexed by historical place names - Time on indexing - Time of using indexed object - Content queried by - contemporary names or - historic names - How to solve the mismatch in information retrieval?
Analysis of change types An example of a change in partonomy: Petsamo was annexed from Finland to USSR on 1944-09-19
Analysis of change types An example of a merge An example of a split Two examples of names changes Region moved Region moved Region moved
Analysis of change types Statistics of different types of regional changes between 1865 and 2007 in Finland
The metadata schema for representing changes
The metadata schema for representing changes Instances are maintained in a spreadsheet application (Excel or OpenOffice Calc)
An example of filled up Metadata Schema of Changes SCHEMA... from = Finland to = USSR movedpart = Petsamo date = 1944-09-19 changetype = changepartof... Knowledge to be modeled
The metadata schema for representing additional properties such as polygonal boundaries
The process Ontology time-series created using schemas and rules Ontology time-series enriched with additional properties such as polygonal data and size data
Finnish Spatio-temporal Ontology SAPO - The metadata schemas and methods were implemented to create an ontology time series of Finnish municipalities between 1865-2007 - Statistics: - RDF repository contains - 1105 different changes - 976 different temporal parts of 616 different historical regions - each has an average 1.58 changes - temporal parts and partonomies constitute 142 different temporal ontologies
Illustration of the global overlap table Shades of grey indicate the level of coverage: the darker the box, the higher is the coverage. The black color indicates a full 100% coverage between the SAPO regions and the white color a 0% coverage. From this illustration it is easy to see the mutual asymmetric coverages between the regions
Evaluation and testing - The system was used in a case study of creating a complete model of the changes of the Finnish municipalities in 1865 2007 - Applied to creating intelligent services and mapbased visualizations in the semantic cultural heritage portal CULTURESAMPO Finnish Culture on the Semantic Web - Applied to the ONKI query expansion widget
Applications - Temporal parts are used to visualize polygonal boundaries of historical regions in CultureSampo and for searching historical artifacts - Items (paintings, photos, etc.) from overlapping regions are listed as recommendations
Applications - Using multiple maps simultanously to visualize relationships between modern regions and historical regions
Application: Query Expansion Widget Using ONKI
Gracias! Kiitos! Thank You!