Ontologies Fmi 042010

Ontologies

Mariana Damova, PhD

April, 2010

Outline

• Definition of ontologies

• History of the science of categorization

• Knowledge models

• Knowledge organization

• Use and Building of ontologies

• Ontology tools

• Ontologies on the Web

Ontologies

What Is An Ontology

• An ontology is an explicit description of a domain:– concepts– properties and attributes of concepts– constraints on properties and attributes– Individuals (often, but not always)

• An ontology defines – a common vocabulary– a shared understanding

Outline






• Ontology tools


Philosophy

• What exists?

• What is?

• What am I?

• What is describing this to me?

Philosophy

Greek etymology

Parmenides of Elea, ancient Greek philosopher (early 5th century BCE)

For never shall this prevail,

that things that are not are

Parmenides made the ontological argument against nothingness, essentially denying the possible existence of a void.

Philosophy

• Jacob Lorhard, German philosopher (1561 - 1609)

• 1607 - First occurrence of the word Ontology (lat. Ontologia) and the first published ontology

Lorhard‘s Ontology• Translation from: Peter Øhrstøm, Sara L. Uckelman; Henrik

Schärfe –• Historical and conceptual foundations of diagrammatical ontology

Ontologies and CS

• Tom Gruber, 1992

• An ontology is a specification of a conceptualization.

• An ontology defines

• Concepts

• Relationships

• Any other distinctions that are relevant for modeling a domain

Ontologies and CS

• To share common understanding of the structure of information among people or software agents

• To enable reuse of domain knowledge

• To make domain assumptions explicit

• To separate domain knowledge from the operational knowledge

• To analyze domain knowledge

Outline






• Ontology tools


Knowledge Models

Structured representations of knowledge using symbols torepresent pieces of knowledge and relationships betweenthem.

Different types of KM have different degrees of formality andlevels of expressivity.

A KM can include:Symbolic character-based languages, such as logicDiagrammatic representations, such as networks and laddersTabular representations, such as matricesStructured text, such as hypertext

Knowledge Models - Design

• Knowledge identification

- What?

• Knowledge specification

- How?

• Knowledge refinement

- Validation

Knowledge Models - Types

• Ladders: hierarchical (tree-like) diagrams

• Tables and Grids: tabular representations

• Network Diagrams: shows nodes connected by arrows

The most complex type of KM

Examples include semantic nets and conceptual graphs

Ladder Model Example – British Royal Family

Tabular Model Example – Stock Markets

Network Diagrams – Semantic Nets

• Nodes in the graph represent concepts

• Arcs represent binary relationships between concepts

• Any characteristic that links two concepts: isA, hasColour, hasAge,

LivesIn, etc.

Note the difference between this structure and the ladders.

Network Diagrams – Conceptual Graphs

• Combination between the existential graphs and Semantic nets

• A conceptual graph consists of:

• Concept nodes – represented as rectangular boxes

• Relations nodes – represented as ovals

• One way connections between the nodes – represented as arrows

• Less intuitive then the Semantic Nets

Nemo the fish lives in water

Outline






• Ontology tools


Knowledge Organization

• Thesaurus

• Taxonomies

• Ontologies

Thesaurus

• Similar with dictionaries

• Provides synonyms and antonyms for words, and not definitions

E.g., WordNet

Taxonomies

• Hierarchical structures

• Subtype-supertype relationships, also called parent-child relationships

• Example: Whale is Mammal; Mammal is an Animal (not all Animals are Mammal, not all mammals are Whales)

Taxonomy Examples

• Taxonomies on the Web– Yahoo! Categories

• Catalogs for on-line shopping– Amazon.com product catalog

• Domain-specific standard terminology– Unified Medical Language System (UMLS)– UNSPSC - terminology for products and services

Yahoo

Amazon

Ontology

• More complexA formal definition of ontologies is provided in [Brewster andWilks, 2009]

O = (C,T,R,A,I,V,≤c, ≤t, σR, σA, IC, IT, IR, IA)

• Whereby:

C – Concepts V – ValuesT – Types ≤ – Partial Order on C and TR – Relations σ – FunctionsA – Attributes I – Partial Instantiation FunctionsI – Instances

Knowledge Organization

Differences in the degree of logical rigor, formality and the potential for reasoning over the data structure

Outline






• Ontology tools


How Ontologies can be used

Ontologies

Software

agentsProblem-

solving

methods

Domain-

independent

applications

DatabasesDeclarestructure

Knowledge

bases

Provide

domaindescription

Types of Ontologies

• Upper Ontology – model of the common objects that are applicable across a wide range of domain ontologies

• Domain Ontology – an ontology developed for a specific domain; conforms to an upper ontology

• Application Ontology – an ontology created for a specific application; may conform to a domain ontology

Types of Ontologies - Examples

Upper Ontologies:• Dublin Core• OpenCyc/ResearchCyc• SUMO• DOLCE• PROTON

• Domain Ontologies:• E-business : Rosetta-Net• Medical: UMLS• Engineering: EngMath

Methods for Building Ontologies

• From scratch – conceptual analysis

• Ontology acquisition

• Building on existing ontologies

Conceptual Schema

Sample Class Hierarchy

Sample Property Hierarchy

Modeling of an organization

DBPedia

http://mappings.dbpedia.org/server/ontology/classes

DBPedia

Proton Ontology

Person in DBPedia

Person in Proton

Outline






• Ontology tools


Protege

Protege

Class and Properties

Instances

TopBraid Composer - Diagram

TopBraid Composer - Graph

TopBraid Composer - Layout

Outline






• Ontology tools


OWL: Summary

The Web Ontology Language (OWL) was developed to provide for more expressive ontologies based on a decidable formal logic.

Three flavours of OWL have been specified: OWL Full for full

expressiveness without guarantees of decidability, OWL DL for a compromise expressiveness within the decidable fragment of Description Logic and OWL Lite as a subset of DL.

OWL provides for additional constructs not present in RDFS to define classes and properties. As a result, OWL is well suited to consistency checking and classification tasks.

OWL Lite

The complete language OWL Full has two sublanguages:

• OWL DL (Description Language)

• supports reasoning applications

• has restrictions on OWL Full constructs

• restrictions make reasoning systems decidable

• OWL Lite

• supports only a subset of OWL Full constructs

• provides a minimal set of features allowing the

development of ontologies without the encoding

of complex semantic relationships

OWL Lite - Classes

OWL classes define basic concepts.

A Simple Named Class is defined as follows:

– <owl:Class rdf:ID=„classname“/>

Ex. <owl:Class rdf:ID=„Restaurant“/>

Predefined OWL Classes (Extreme classes)

– Thing class (owl:Thing)

the most general class

every individual is member of this class

– Nothing class (owl:Nothing)

empty class with no member individuals

OWL Lite - Properties

There are four disjoint type of properties in OWL.

• Datatype properties (owl:DatatypeProperty)

• Object properties (owl:ObjectProperty)

• Annotationproperties (owl:AnnotationProperty)

• Ontology properties (owl:OntologyProperty)

OWL Lite – Classes and Properties

Linking Open Data

LDSR – reason-able view to LOD

http://ldsr.ontotext.com

http://linkedlifedata.com

Summary






• Ontology tools


Documents

Ontologies Fmi 042010