It ti Otl Mdl d Integrating Ontology Models and Conceptual ... · • No consistency problem on meta model side • Not all ontology concepts required on meta model side • Consistency

1

I t ti O t l M d l d

Protégé Conference Amsterdam – June 2009

Integrating Ontology Models and Conceptual Models

using a Meta Modeling Approach

Hans-Georg Fill, Patrik BurzynskiDepartment of Knowledge and Business Engineering,

http://www.dke.univie.ac.athttp://www.dke.univie.ac.at

University of Vienna

Agenda

• Motivation• Conceptual Modeling• Conceptual Modeling• Meta Modeling• Three approaches for an integration• Implementation and Application Scenarios• Outlook


2

Motivation

• Several business cases for the use of web-based ontologies:– Enterprise Content Management– Enterprise Information Integration– Enterprise Service Bus– …

• Some common goals:– Let machines manage complexity


– Using explicit semantics and reasoning

– Based on shared, web-based,explicit conceptualizations

• How to analyze the contribution to business value?Picture source: http://www.atibatechnology.com/ECM_DM.html

Conceptual Modeling• Conceptual Modeling:

– Supporting humans to analyze organizational and technical relationshipsp

– Based on formal syntax and notation– Optional definition of formal semantics– Goal: Improvement of Human

understanding

• Examples:– Analysis of business models, strategic goals, performance

i i f i


measurements, opportunities for action– Management of business processes including representation,

analysis, simulation– …

• How to bring together ontologies and conceptual models?

3

BusinessView

Business Processes

ManagementApproaches

eg. BSC

RegulatoryFrameworks

Business Model

IT-GovernanceFrameworks

Production Model

SemiformalProzeßstart-46430

Aktivität-46433

Name

an

Name

an

Percentage ofemployees

with a flexible Strategic

J

Operatives

J

Kennzahl-42841

Kennzahl-42841

J

Business & IT

SemiformalView

Aktivität-46436

Aktivität-46436-46439

Entscheidung-46442

Aktivität-46451

Aktivität-46454

Aktivität-46457

Name

an

Name

anan

Name

an

Name

an

Name

an

an

Name

an

with a flexibletime schedule

x

/

1100,000000

1 2α

Number ofemployees

with a flexibletime schedule

db

αNumber ofemployees

with atraditional

time schedule

db

StrategicGoal

OperativesZiel-42838 4284142841

StrategischesZiel-42844

J

FormalizedViewB-XML

(e.g. BPEL) S-XML C-XML ……

Technology


InfrastructureView

TechnologyView

Services Orchestrated Services Applications… …

Spectrum of „Ontologies“

strongsemantics

First Order LogicModal Logic

Humans usually do not express themselves in a logic-based language…

Thesaurus

Conceptual Model

Logical Theory

ER

Extended ERXTM

RDF/S

UML

DAML+OIL,OWL

Description Logic

Has Narrower Meaning Than

Is Subclass of


weak semantics

TaxonomyRelational Model

Schema

Is Sub-Classification of

(Obrst, 2003)

“Ontologies are usually expressed in a logic-based language…”

4

DKE Meta Modeling Framework

modelingmethod

defines grammar

defines meaningModellingLanguagemodelinglanguage

defines visualizationdefines way of language application

deliversmodelingprocedure

modelingtechnique

method

mechanisms& algorithms

used for

used in

genericmechanisms& algorithms

specific


arrangesaccording to

Semantics

SemanticSchema

Syntax

SemanticMapping

connectsconsiders

Notation semantics

semanticdomain

syntaxnotation visualizes

semanticmapping

describesmeaning of

steps results

hybridmechanisms& algorithms

mechanisms& algorithms

(Karagiannis et al., 2002)

Meta Modeling Approach

0..* 1..1Ontological VisualizationPattern

External Variable0..* 1..1Ontological Visualization

PatternExternal Variable

Ontological VisualizationPattern

External Variable

Visualization (Notation)Meta Model (Syntax)

BusinessProcess

Flow Element ControlC t

IsFromActivityElement

WorkingEnvironment

WE Element WE Relation

IsFromWEElement

Business Graph

Relations

Personibl f

RequiredR

1..* 1..1

1..1

1..* 1..*

1..1

1..1

1..1

1..*

1..1

1..*

1..1

1..* 1..*

1..* 1..1

1..1 1..1

1..1

1..* 1..*

1..1

0..*0..*

1..*

1..* 0..*

1..*

0..*

1..*consists of contains

linked to

refers to

Visual Object

Graphical Primitive Internal Variable

Transformation Function

refers toconsists of

1..1

1..* 1..*

1..1

0..*0..*

1..*

1..* 0..*

1..*

0..*

1..*consists of contains

linked to

refers to

Visual Object


Transformation FunctionVisual Object


Transformation Function

refers toconsists of

Connector

Activity SubProcessPerson Group Resource

responsible for Resource

responsibleFor

responsibilityBy

requiredFor provisionedBy

IsToActivityElement IsToWEElement1..* 1..11..* 1..1

1..* 1..1

1..11..*

1..*

1..*

1..*

1..*

Control ElementControl Element

StartStart DecisionDecisionEndEnd

Identification of Common Model Characteristics


5

Example for the Realization of a Process Meta Model

S

Class Relationclass

C t t

Notify Customer

Customer reached

ActivityDecision

Subsequent

Meta ModelModel


Contact Customer

Write E-MailCustomer not reached

Semantic Issues

• Meta Models for Conceptual Modeling:– Assumption of implicit human-interpreted semanticsAssumption of implicit, human interpreted semantics– Only „execution“ of models requires strict formalization incl.

formal semantics– Formalization of semantics only for particular purposes, e.g.

simulation– No consideration of inherent semantics of the model content,

e.g. an activity in a process is named „print report“ but no information about print‘ or report‘ is made explicit


information about ‚print or ‚report is made explicit– IT-based communication of semantic information requires

common semantic base, e.g. an ontology

How to make semantics explicit?

6

Integrating Explicit Semantics

Three Approaches:Three Approaches:• Integrating Ontologies on the Meta Model Level• Defining References between Meta Model Elements

and externally kept Ontologies, resp. between Model Elements and Ontologies

• Using a Combination of Integration and External Li k


Linkage

Example for the Realization of an Ontology Model

O C

Class Attribute

OWL ClassProperty

Domain

Meta ModelModel

C C C C


foaf:Document dcmitype:Image skos:ConceptScheme skos:Collection

PaltLabel

PaltSymbol

PchangeNote

Pdc:title

7

Integrating Ontologies on the Meta Model Level

Linkages on the Meta Model LevelMeta Model Ontology Meta Model

Contact Customer

Notify Customer

Write E-Mail

Customer reached

Customer t h d

ActivityDecision

Subsequent

Class Relationclass

Meta Model

Model

OWL ClassProperty

Domain

Class Attribute

Meta Model

Model

Cfoaf:Document

Cdcmitype:Image

Cskos:ConceptScheme

Cskos:Collection

P P P P

Meta Model Ontology Meta Model


not reachedaltLabel altSymbol changeNote dc:title

Advantages:• Coherent conceptual approach for both ontologies and domain aspects• Inherent graphical representation of ontology• General mechanisms and algorithms for meta models directly applicable to ontologies

Disadvantages:• Graphical modeling of large ontologies is difficult• Visualization of large and complex ontologies difficult to handle for users• Specific ontology aspects might need to be re-implemented / re-configured

References between (Meta) Models and External Ontologies

Meta Model Ontology Meta Model

Contact Customer

Notify Customer

Write E-Mail

Customer reached

Customer not reached

ActivityDecision

Subsequent

Class Relationclass

Meta Model

Model

Extensions Repository

Class

Attribute

Contact Customer

C


Advantages:• Independent modifications of ontologies in the repository possible, e.g. sharing with third parties• Ontology repository may better meet user needs for ontology manipulation• Less implementation effort on meta model side

Disadvantages:• Consistency problems, e.g. in case of modifications of the ontology or the models• Different representations of models and ontologies may require extra user training• Possible time delays in interaction btw. systems

8

Using a Hybrid Approach of Integration and Referencing

Class Relationclass

Ontology Repository

Meta Model

Class

Ontology Meta Model

Contact Customer

Notify Customer

Write E-Mail

Customer reached


ActivityDecision

Subsequent

Meta Model

Model

Advantages: Disadvantages:

OWL ClassProperty

Cskos:Collection

Cskos:OrderedCollection

Pdc:title

Pdefinition


• Coherent conceptual approach for both ontologies and domain aspects• Independent modifications of ontologies in the repository possible, e.g. sharing with third parties• Ontology repository may better meet user needs for ontology manipulation• No consistency problem on meta model side• Not all ontology concepts required on meta model side

• Consistency problems still persist btw. ontology meta model and ontology repository• Specific ontology aspects might need to be re-implemented on meta model side

The meta modeling approach has been implemented as a multi-user plattform (ADONIS®).

Implementation Platform

a multi user plattform (ADONIS ).

External interfaces

DocumentGeneration

Process Cost

Component

Transformation

Import/Export

Acquisition Modeling Analysis Simulation Evaluation

User Interaction Adonis CommunityEdition for freedownload:

http://www.adonis-community com


Database

CORE (Modeling subsystem)

Component community.com

9

Scenarios

Scenario A: AGES Management of Clinical Trials→ Integrating Ontologies on the Meta Model Level→ Integrating Ontologies on the Meta Model Level

Scenario B: Semantic Culture Guide→ Using a Hybrid Approach of Integration and

Referencing via Protégé


Implementation A

ActivityDecision

Subsequent

Class Relationclass

Meta Model Term Meta Model

Term

Class

Contact Customer

Notify Customer

Write E-Mail

Customer reached


DecisionMeta Model

ModelMeta ModelModel

WIN/NATAallgemein

Twenn Antragunvollständig

Twenn Antragvollständig

TFachliche

Beurteilung

TNegative

Beurteilung -Antrag hat Mängel

T

keineBeseitigung der

Mängel -Fristablauf

TPositive

Beurteilung -keine Mängel

T


Visual views on process models based on terms

10

Goal: „Make Austria’s culture events accessible through a decentralized one-stop shop by using semantic technologies.”

SCG Project Outline

Innovative Aspects:

• Creation of an adequate Modeling Method for the description of Business Models, Business Processes and Services of the organizers of culture events.

• Creation of a Culture-Ontology for the coherent description of the involved actors, the culture events and genres and the user feedback dimensions


• Implementation of semantic services and serviceworkflows for the support of visitors and event organizers

• Practical trial of the semantic services and workflows

Implementation B

A ti it S b t

Class Relationclass

Ontology Repository

Meta Model

Class

Ontology Meta Model

Contact Customer

Notify Customer

Write E-Mail

Customer reached


ActivityDecision

Subsequent

Meta Model

Model

OWL ClassProperty

Cskos:Collection

Cskos:OrderedCollection

Pdc:title

Pdefinition

+ ADONIS Export Plugin


11

Outlook: www.openmodels.at

Open Model Community

Open Model Projects

Open Model Technologies


Community Projects Technologies

SeMFIS on Open Model

• Semantic-based Modeling Framework for Information Systems (SeMFIS)

• Provision of modeling framework, technologies, and tools to support semantic information models

• Current tasks:– Provision of a web-based modeling tool based on Java

applets (AdoWeb)– Coupling of AdoWeb and Protégé on a common platform


– Extension of the modeling functionalities for using ontologies and conceptual models

– …

12

SeMFIS Community


SeMFIS Technologies: Web-based Modeller


13

SeMFIS Technologies: Applet Version of Protégé incl. Plugin


Thank you for your attention!


y [email protected]

Documents

It ti Otl Mdl d Integrating Ontology Models and Conceptual ... · • No consistency problem on meta model side • Not all ontology concepts required on meta model side • Consistency