120 likes | 221 Views
Second International Workshop on New Generation Enterprise and Business Innovation NGEBIS 2013. Linked Data Based Approach to Similarity Reasoning Anna Formica, Michele Missikoff , Elaheh Pourabbas , Francesco Taglino CNR-IASI, Italy. Outline. Motivations & Objectives
E N D
Second International Workshop on New Generation Enterprise and Business InnovationNGEBIS 2013 Linked Data Based Approach to Similarity Reasoning Anna Formica, Michele Missikoff, ElahehPourabbas, Francesco Taglino CNR-IASI, Italy
Outline • Motivations & Objectives • Ourproposal • Linked Data principles • SemSim in a nutshell and namespace • Conclusions
Motivations and Objectives • Knowledge as a key fertilizer to allow an innovative idea to grow into a concrete value proposition for enterprises • To overcome the difficulty in finding the knowledge that appears relevant to the proposed innovative idea • To concretely extract and use the identified knowledge resources, despite of the different formats and applications used to generate them.
OurProposal • SemSim: a Semantic Similarity reasoning method to retrieve semantically annotated knowledge resources • Representation of the SemSim notions according to the Linked Data principles to enhance • interoperability • Semantic Web standards engagement • Reuse of existing resources
Linked Data and … A set ofprinciplestoallow • Standard descriptionof data (RDF-based) • Stardard way ofaccessing data (HTTP) • Linkingresources/data amongthem • Goingbeyondbypublishingservices in a Linked Data fashion
SemSim in a nutshell • Weighted Reference Taxonomy (WRT) • concepts reports the probability to describe a resource • Ontology Feature Vectors (OFV) to express • Semantic annotation of digital resources • User requests • Semantic Similarity (semsim function) • Based on similarity between concepts: consim • Results as a ranked list of resources according to their similarity wrt the request
WRT and OFVs OFV forannotatingresources dr1 = (InternationalHotel, FrenchMeal, Cinema, Flight) • dr2 = (Pension, VegetarianMeal, ArtGallery, ShoppingCenter) • dr3 = (CountryResort, MediterraneaMeal, Bus) • … • dr7 = (RegularAccommodation, RegularMeal, Salon, Flight) • … • dr15 =(InternationalHotel, PictureGallery, Flight) • …. OFV forrepresenting a UserRequest rv = (InternationalHotel, InternationalMeal, ArtGallery, Flight)
SemSim in RDF syntax • SKOS torepresent the Taxonomy • semsimnamespacedefining the SemSimvocabulary
SemSimnamespace OFV rdf:typeowl:Class. AnnotVrdf:typeowl:Class; rdfs:subClassOf OFV. RV rdf:typeowl:Class; rdfs:subClassOf OFV. RSV rdf:typeowl:Class. hasWeightrdf:typeowl:DatatypeProperty; rdfs:domainskos:Concept; rdfs:rangexsd:float. hasConceptrdf:typeowl:ObjectProperty; rdfs:domain OFV; rdfs:rangeskos:Concept. refersTordf:typeowl:ObjectProperty; rdfs:domainAnnotV. hasScorerdf:typeowl:DatatypeProperty; rdfs:domainRSVElem; rdfs:rangexsd:float hasAnnotVrdf:typeowl:ObjectProperty; rdfs:domainRSVElem; rdfs:rangeAnnotV. hasRSVElemrdf:typeowl:ObjectProperty; rdfs:domain RSV; rdfs:rangeRSVElem.
Conclusions and Outlook • Representing SemSim notions in a RDF syntax to publish SemSim service in a Linked Data • From Linked Data to Linked Services