OWL (Ontology Web Language and Applications) Maw-Sheng Horng Department of Mathematics and Information Education National Taipei University of Education

Content -Origins of OWL -Purposes of OWL -Three Different Species of OWL -Class Hierarchies -Using OWL to Define Classes -OWL: Thing -Using OWL to Define Properties -Syntax for indicating that a property is symmetric -Reasoning on the Web -SPARQL: a Simple Protocol And Resource description framework Query Language - Some other OWL Tools -Some Typical OWL Applications - Conclusions

Origins of OWL DAML DAML+OIL DAML = DARPA Agent Markup Language OIL = Ontology Inference Layer RDF = Resource Description Framework OWL is now on track to become a W3C Recommendation! OIL OWL RDF All were influenced by RDF

OWL and RDF Schema enables machine-processable semantics XML/DTD/XML Schemas RDF Schema OWL Syntax Semantics OWL= RDFS +more

Purposes of OWL The purpose of OWL is identical to RDF Schemas - to provide an XML vocabulary to define classes, their properties and their relationships among classes. –RDF Schema enables you to express very rudimentary relationships and has limited inferencing capability. –OWL enables you to express much richer relationships, thus yielding a much enhanced inferencing capability. A benefit of OWL is that it facilitates a much greater degree of inference making than you get with RDF Schemas.

Purposes of OWL OWL gives you an XML syntax to express statements about properties and classes, above and beyond what you can make with RDF Schema. –Using OWL to define classes. –Using OWL to define properties. –Using OWL to define relationships. –OWL statements that you can incorporate into your instance documents.

Three Different Species of OWL OWL Lite : based on a fragment of FOL (First-Order Logic) OWL DL (Description Language): based on the description logic SHION, a larger fragment of FOL, decidable for many applications, but it has a rather high worst-case time complexity (Exponential Time) OWL Full: combination of OWL DL with all features of RDFS, and undecidability

Class Hierarchies Class1 Class2Class3 Properties: property1: Type1 property2: Type2... These are the properties of Class1. The name of the property is shown (e.g., property1), and its range is shown in italics (e.g., Type1). Class2 and Class3 are subclasses of Class1. Class1 Class2 Class3 Properties: property1: Type1 property2: Type2... An alternate notation to the above class hierarchy is to use a Venn diagram, as shown here. Taxonomy (Class Hierarchy): Venn Diagram:

Using OWL to Define Classes OWL gives you the ability to construct classes using the following set operators: -IntersectionOf -UnionOf -ComplementOf -Construct a class by enumerating it instances -Specify a class that is equivalent to (or disjoint from) another class

OWL: Thing OWL: Thing is a predefined OWL class. All instances are members of owl:Thing, Thing class is the root of all classes.

Using OWL to Define Properties Property Characteristics: -Range: indicate the possible values for a property -Domain: associate a property with a class -subPropertyOf: specialize a property -Symmetric Property -Transitive Property -Functional Property -Inverse Property -ObjectProperty: relates a resource to another resource -DataProperty: relates a resource to a Literal or an XML schema datatype

<rdf:RDF xmlns:rdf=" xmlns:rdfs=" xmlns:owl=" xml:base=" Syntax for indicating that a property is symmetric

Reasoning on the Web -Class membership: If x is an instance of class c, and c is a subclass of D, then we can infer that x is an instance of D. - Equivalence of classes: if c is equivalent to D, and D to E, then C is equivalent to E. -Consistency(concept satisfiability): if x is an instance of C, and if C is a subclass of intersection of D and E, with D and F are incompatible, then there must exist an inconsistency, and the class C shoud be empty. -Classification(instance checking): if certain property-value pairs have been declared as sufficient conditions for membership to a class C, then if an individual x satisfies such conditions, it can be concludes that x is an instance of C.

-Concept subsumption: is concept C more general than D? -Knowledge base satisfiability: is the combined knowledge of Tbox (Terminology box) and Abox (Assertion box) free of contradictions? -Knowledge correctness : if knowledge base are sound and complete? -Some reasoners: *KAON *Pellet *FACT * RacePro

SPARQL: a Simple Protocol And Resource description framework Query Language Three alternative kinds of query: -ASK: Tests query pattern, return yes/no result. -CONSTRUCT : Extract part or whole graphs from knowledge base, Transforming from one representation into another. -DESCRIBE: Return RDF graph that describes set of resources, useful for constructing RDF graphs from non-RDF data sources.

-D2RQ: A server converts relational database to RDF. -CORESE: A search engine for RDF dataset and provides SPARQL interface. -Jena: A semantic web framework for JAVA environment. -SeWeSe: A semantic web server 1 Some other OWL Tools

Some Typical OWL Applications: -Social network analysis: degree centrality, betweenness centrality -FOAF (Friends of A Friend) ontology -Semantic WIKI -Relational databases semantic integration -Web services integration

Conclusions Advantages of using OWL to define an ontology: -Extensible: much easier to add new properties -Portable: much easier to move an OWL document than to move a database -Reasoning support: can generate implicit data from explicit data -Web 3.0 foundation