a method for reusing and re-engineering non-ontological resources for building ontologies

A Method for Reusing and Re-engineering Non-ontological

Resources for Building Ontologies

Boris Villazón-Terrazas

[email protected]

th14th April, 2011

Outline

2

Introduction

NACE

3

GoalIntroduction

ToTo speedspeed up up thethe ontologyontologydevelopmentdevelopment byby reusingreusing

and reand re‐‐engineeringengineeringnonnon‐‐ontologicalontological resourcesresourcesthat have already reached that have already reached

bbsome consensus by some consensus by standardization bodiesstandardization bodies

4

From scratchOverview – Methodologies for building ontologies

Introduction

From scratch• METHONTOLOGY [Gómez-Pérez et al. 2003]• On-To-Knowledge [Staab et al. 2001]• DILIGENT [Pinto et al. 2004]

NeOn Methodology [Suárez-Figueroa 2010]NeOn Methodology [Suárez Figueroa 2010]

5

The role of PatternsIntroduction

Architecture [Alexander 1979]

Software engineering [Tichy 1997]

Knowledge engineering [Clark et al 2000]Knowledge engineering [Clark et al. 2000]

Ontology engineeringgy g g• Semantic-Web Best Practices and Deployment Working Group• Ontology Design Patterns Public Catalog• Ontology Design Patterns Portal• Ontology Design Patterns Portal• Linked Data Patterns

6

Apply software re-engineering concepts to Ontology EngineeringIntroduction

SoftwareDevelopment

P S

Software A

SoftwareSoftwareProcess Save resources

Save time

Software B

Software Re‐

EngineeringEngineering

Save timeengineeringPatterns

Ontologyl

Save resources

Classification Scheme

Thesaurus

Lexicon

GlossaryOntologyOntologyDevelopment

Process

Save time

y

FolksonomyEngineeringEngineering

7

Reach consensus

Ontology Re‐engineeringPatterns

Outline

8

State of the Art

9

Software re-engineering modelState of the Art

Forward Con-

(Alteration)

Reverse Engineering

Con-

t lre-think

Engineering

(Refinement)

Design

ceptual

Requirements

Engineering(Abstraction)

ceptual

Requirements

Designre-design

re-specify

Implementation

Design Design

Implementationre-code

re-design

Existing System Target System

Software re-engineering Patterns describe how to change a legacy system into a new,refactored system that fits current conditions and requirements [Pooley et al. 1998]

10

E.J. Byrne. A conceptual foundation for software re-engineering. In Proceedings of the International Conference on Software Maintenance and Reengineering, pages 226–235. IEEE Computer Society, 1992.

Ontology PatternsState of the Art

Online catalogues

SemanticWeb Best Practices and

DeploymentWorkingGroup

Ontology DesignPatterns Public

Catalog

Ontology DesignPatterns (ODP)

PortalLinked Data Patterns

L k f tt f i i i t t l i

11

Lack of patterns for re-engineering resources into ontologieshttp://ontologydesignpatterns.org

State of the Art

12

Evaluation frameworkNon ontological resource

State of the Art

• type• data model• implementation

• type• data model• implementation

Non-ontological resource

• methodological guidelines• tool support• provenance

• methodological guidelines• tool support• provenance

Reuse

• provenance• provenance

• approach: ABox,TBox or Populationsemantic/s ntactic le el

• approach: ABox,TBox or Populationsemantic/s ntactic le el

Transformation

• semantic/syntactic level• makes explicit the semantics hidden in the relations• relies on additional resources or human domain expert• automatic / semiautomatic / manual• methodological guidelines

t h i l d

• semantic/syntactic level• makes explicit the semantics hidden in the relations• relies on additional resources or human domain expert• automatic / semiautomatic / manual• methodological guidelines

t h i l d• techniques employed• techniques employed

• lightweight or heavyweight• lightweight or heavyweight

Ontology

13

• implementation language• single or several• implementation language• single or several

18 methods & 13 toolsState of the Art

Non-ontological resource Reuse

Research work

Type of resource

Data model is known

Resource

implemented in Hepp et al. [HdB07] Classification

scheme, thesauri

No Database

Hakkarainen et al. [HHST06]

Classification scheme

Yes Flat file

Abbasi et al. [ASC07] Folksonomy No Not mentionedMaala et al. [MDA07] Folksonomy No Not mentionedvan Assem et al. Lexica Yes Prolog

Research work Methodological Guidelines

Tool Support

Provenance

Hepp et al. [HdB07] No No NoHakkarainen et al. [HHST06]

No No No

Abbasi et al. [ASC07] No No NoMaala et al. [MDA07] No No NoVan Assem et al. [vAGS06]

No No Yes

[vAGS06] Gangemi et al. [GNV03, GGMO03]

Lexica Yes Database

Hahn et al. [Hah03, HS03] Thesauri Yes ASCII filesVan Assem et al. [vAMSW04]

Thesauri No proprietary text format, database, XML

Van Assem et al. [vAMMS06]

Thesauri No Not mentioned

Wielinga et al. [WSWS01] Thesauri Yes XMLHyvo nen et al. [HVTS08] Thesauri No Not mentionedSoergel et al. [SLL+ 04, LS06]

Thesauri Yes Database

Gangemi et al. [GNV03, GGMO03]

No No No

Hahn et al. [Hah03, HS03] No No NoVan Assem et al. [vAMSW04]

No No No

van Assem et al. [vAMMS06]

No No No

Wielinga et al. [WSWS01] No No NoHyvo nen et al. [HVTS08] No No NoSoergel et al. [SLL+04, LS06]

No No No

Stojanovic et al. [SSV02] Not specified Yes DatabaseBarrasa et al. [BCGP04, Bar07]

Not specified Yes Database

Garcıa et al. [GC05] Not specified No XMLAn et al. [AM05] Not specified No XMLCruz et al. [CXH04] Not specified No XMLFoxvog et al. [FB06] Not specified No Flat file

Stojanovic et al. [SSV02] No No NoBarrasa et al. [BCGP04, Bar07]

No No No

Garcıa et al. [GC05] No No NoAn et al. [AM05] No No NoCruz et al. [CXH04] No No NoFoxvog et al. [FB06] No No No

T f i O t lResearch work Transformat

Approach Transformation

Aspects Semantics of

NOR relations Additional

Resources /

Domain

Expert

Automatic /

Semi-

automatic /

Manual

Methodological

Guidelines Technique Tool support

Hepp et al. [HdB07] TBox syntactic, semantic

subClassOf, ad-hoc relation

No Semi-automatic Yes Not mentioned SKOS2GenTax


ABox syntactic, semantic


No Semi-automatic Yes Transformation rules

Not mentioned

Abbasi et al. [ASC07] Population syntactic No Swoogle Google

Automatic Yes Lexico Syntactic Patterns

T-ORG

Maala et al. [MDA07] Population syntactic No WordNet, G hi l

Automatic Yes Not mentioned Not mentioned

TransformationResearch Work

Lightweight/ Heavyweight

Components Implementation language

Single/ Several

Hepp et al. [HdB07] Lightweight classes, relations RDF(S) / OWL DLP

Single


Lightweight classes, attributes, relations

OWL DL Single

Abbasi et al. [ASC07] Lightweight instances Not mentioned SeveralMaala et al. [MDA07] Lightweight instances RDF SingleVan Assem et al. [vAGS06]

Lightweight classes, attributes, relations, instances

RDF(S) / OWL Full

Single

Gangemi et al [GNV03 Lightweight classes attributes DAML+OIL Single

Ontology

Geographical locations,

Ontology of things

Van Assem et al. [vAGS06]

ABox syntactic No No Semi-automatic Yes Not mentioned Swi-Prolog


TBox syntactic, semantic

ad-hoc relations DOLCE Semi-automatic Yes NLP Techniques

Not mentioned

Hahn et al. [Hah03, HS03]


subClassOf, partOf, ad-hoc

relation No Semi-automatic Yes Ontology

Design Patterns

Ad-hoc tool

van Assem et al. [vAMSW04]



No Semi-automatic Yes Not mentioned Ad-hoc tool


Population syntactic Not mentioned No Automatic Yes Not mentioned Swi-Prolog

Wielinga et al. [WSWS01]


subClassOf Not mentioned Semi-automatic Yes Not mentioned Ad-hoc tool

Hyvo nen et al. TBox syntactic, subClassOf, DOLCE Semi-automatic Yes Not mentioned Ad-hoc tool



DAML+OIL Single


Heavyweight classes, relations LOOM / ALC Single



RDF(S) / OWL Full

Single


Lightweight instances SKOS RDF Single



RDF(S) Single

Hyvo nen et al. [HVTS08]

Heavyweight classes, attributes, relations

RDF(S) Single

Soergel et al. [SLL+04, LS06]


OWL DL Single

Stojanovic et al. Lightweight instances F-Logic / RDF Single

14

y[HVTS08]

y ,semantic

,partOf

Soergel et al. [SLL+ 04, LS06]



No Manual Yes Not mentioned Not mentioned

Stojanovic et al. [SSV02]

Population syntactic ad-hoc relation No Semi-automatic Yes Mapping rules KAON- REVERSE

Barrasa et al. [BCGP04, Bar07]

Population syntactic subClassOf, ad-hoc relation

No Semi-automatic Yes Mapping rules ODEMapster

Garcıa et al. [GC05] ABox syntactic ad-hoc relation No Semi-automatic Yes Mapping rules XSD2OWL XML2RDF

An et al. [AM05] ABox syntactic ad-hoc relation No Semi-automatic No Not mentioned Discovery tool

Cruz et al. [CXH04] ABox syntactic Not mentioned No Semi-automatic Yes Mapping rules Ad-hoc tool

Foxvog et al. [FB06] ABox syntactic Not mentioned No Semi-automatic Yes Not mentioned Ad-hoc tool


Lightweight instances F Logic / RDF Single


Lightweight instances RDF Single

Garcıa et al. [GC05] Lightweight classes, attributes, relations, instances

OWL Full/ RDF Single

An et al. [AM05] Lightweight instances RDF SingleCruz et al. [CXH04] Lightweight classes, attributes,

relations, instances RDF(S) Single

Foxvog et al. [FB06] Lightweight classes, attributes, relations, instances

CycL / OWL Full / WSML

Several




Tool Support

Provenance


No No No


No No Yes

Research work

Type of resource

Data model is known

Resource

implemented in Hepp et al. [HdB07] Classification

scheme, thesauri

No Database

Tool

Type of resource

Data model model is known

Resource implemented in

SKOS2GenTax Classification h

No SKOS RDF Gangemi et al. [GNV03, GGMO03]

No No No


No No No


No No No


No No No

sauri Hakkarainen et al. [HHST06]

Classification scheme

Yes Flat file

Abbasi et al. [ASC07] Folksonomy No Not mentionedMaala et al. [MDA07] Folksonomy No Not mentionedvan Assem et al. Lexica Yes Prolog

schemes, thesauri

T-ORG Folksonomy No Not mentioned KAON- REVERSE



No No No


T f i O t l

[vAGS06] Gangemi et al. [GNV03, GGMO03]

Lexica Yes Database

Hahn et al. [Hah03, HS03] Thesauri Yes ASCII filesVan Assem et al. [ AMSW04]

Thesauri No proprietary text f t d t b

ODEMapster Not specified Yes DatabaseD2R Server Not specified No DatabaseTopBraid Com- poser

Not specified No Database, XML, Flat file, Spread- sheet

Research work Transformat




Resources /

Domain

Expert

Automatic /

Semi-

automatic /

Manual

Methodological









Not mentioned



T-ORG






Single/ Several


Single



OWL DL Single



RDF(S) / OWL Full

Single


Ontology[vAMSW04] format, database, XML

Van Assem et al. [vAMMS06]

Thesauri No Not mentioned

Wielinga et al. [WSWS01] Thesauri Yes XMLHyvo nen et al. [HVTS08] Thesauri No Not mentioned

sheetXSD2OWL and XML2RDF

Not specified No XML

ConvertToRdf Not specified No Delimited text data file


Ontology of things






Not mentioned





Design Patterns

Ad-hoc tool













DAML+OIL Single





RDF(S) / OWL Full

Single





RDF(S) Single



RDF(S) Single



OWL DL Single


y [ ]Soergel et al. [SLL+ 04, LS06]

Thesauri Yes Database

Stojanovic et al. [SSV02] Not specified Yes DatabaseBarrasa et al. [BCGP04, Bar07]


flat2rdf Not specified No Flat fileJava BibTeX-To- RDF Converter

Not specified No Flat file

Excel2rdf Not specified No Spreadsheet RDF123 Not specified No Spreadsheet

15

y[HVTS08]

y ,semantic

,partOf
























Several

Garcıa et al. [GC05] Not specified No XMLAn et al. [AM05] Not specified No XMLCruz et al. [CXH04] Not specified No XMLFoxvog et al. [FB06] Not specified No Flat file

RDF123 Not specified No Spreadsheet XLWrap Not specified No Spreadsheet


ReuseNon-ontological resource


Tool Support

Provenance


No No No


No No Yes

• Ad-hoc solutions for particular resources• Only focus on the resource type or

i l t ti Gangemi et al. [GNV03, GGMO03]

No No No


No No No


No No No


No No No

resource implementation


No No No






Resources /

Domain

Expert

Automatic /

Semi-

automatic /

Manual

Methodological









Not mentioned



T-ORG






Single/ Several


Single



OWL DL Single



RDF(S) / OWL Full

Single


Ontology


Ontology of things






Not mentioned





Design Patterns

Ad-hoc tool













DAML+OIL Single





RDF(S) / OWL Full

Single





RDF(S) Single



RDF(S) Single



OWL DL Single


16

y[HVTS08]

y ,semantic

,partOf
























Several



• Ad-hoc solutions• Only focus on the resource type or

i l t ti


Tool Support

Provenance


No No No

resource implementation Abbasi et al. [ASC07] No No NoMaala et al. [MDA07] No No NoVan Assem et al. [vAGS06]

No No Yes

Gangemi et al. [GNV03, No No No

T f i O t l


No No No


No No No

van Assem et al No No No





Resources /

Domain

Expert

Automatic /

Semi-

automatic /

Manual

Methodological









Not mentioned



T-ORG






Single/ Several


Single



OWL DL Single



RDF(S) / OWL Full

Single


Ontologyvan Assem et al. [vAMMS06]

No No No


No No NoGeographical

locations, Ontology of

things Van Assem et

al. [vAGS06] ABox syntactic No No Semi-automatic Yes Not mentioned Swi-Prolog




Not mentioned





Design Patterns

Ad-hoc tool













DAML+OIL Single





RDF(S) / OWL Full

Single





RDF(S) Single



RDF(S) Single



OWL DL Single


LS06]Stojanovic et al. [SSV02] No No NoBarrasa et al. [BCGP04, Bar07]

No No No

Garcıa et al. [GC05] No No NoA l [AM05] N N N

17

y[HVTS08]

y ,semantic

,partOf
























Several

An et al. [AM05] No No NoCruz et al. [CXH04] No No NoFoxvog et al. [FB06] No No No



• Ad-hoc solutions solutions for particular resources• Only focus on the resource type or

i l t ti

• No detailed guidelines for the selection of theresource are given

Th i f ti i tresource implementation •The resource provenance information is not •considered





Resources /

Domain

Expert

Automatic /

Semi-

automatic /

Manual

Methodological









Not mentioned



T-ORG






Single/ Several


Single



OWL DL Single



RDF(S) / OWL Full

Single


Ontology


Ontology of things






Not mentioned





Design Patterns

Ad-hoc tool













DAML+OIL Single





RDF(S) / OWL Full

Single





RDF(S) Single



RDF(S) Single



OWL DL Single


18

y[HVTS08]

y ,semantic

,partOf
























Several







Resources /

Domain

Expert

Automatic /

Semi-

automatic /

Manual

Methodological


Hepp et al. [HdB07] TBox syntactic, subClassOf, No Semi-automatic Yes Not mentioned SKOS2GenTax

• ad-hoc solutions• only focus on the resource type or

i l t ti

• no detailed guidelines for the selection of theresourced t id th

semantic ad-hoc relation





Not mentioned



T-ORG

Maala et al. [MDA07] Population syntactic No WordNet, Automatic Yes Not mentioned Not mentioned

Tool Transformat




Resources /

Domain

Automatic /

Semi-

automatic /

Technique

resource implementation • do not consider the resource provenance information

[ ] p y ,Geographical

locations, Ontology of

things Van Assem et

al. [vAGS06] ABox syntactic No No Semi-automatic Yes Not mentioned Swi-Prolog

Gangemi et al. TBox syntactic, ad-hoc relations DOLCE Semi-automatic Yes NLP Not mentioned

Domain

Expert automatic /

Manual SKOS2GenTax TBox syntactic,

semantic subClassOf,

ad-hoc relation No Semi-automatic Not mentioned

T-ORG Population syntactic No Swoogle Automatic Lexico Syntactic

O t l

[GNV03, GGMO03] semantic Techniques





Design Patterns

Ad-hoc tool





van Assem et al. Population syntactic Not mentioned No Automatic Yes Not mentioned Swi-Prolog

Patterns

KAON-REVERSE ABox syntactic No No Semi-automatic Mapping rules

ODEMapster Population syntactic No No Semi-automatic Mapping rules

D2R Server Population syntactic No No Semi-automatic Mapping rules

TopBraid Composer ABox syntactic No No Semi-automatic Mapping rules

Research Work Lightweight/ Heavyweight


Single/ Several


Single



OWL DL Single



RDF(S) / OWL Full

Single


Ontologyvan Assem et al. [vAMMS06]

Population syntactic Not mentioned No Automatic Yes Not mentioned Swi Prolog






subClassOf, partOf

DOLCE Semi-automatic Yes Not mentioned Ad-hoc tool

Soergel et al. [SLL+04 LS06]


subClassOf, ad hoc relation


p p y pp g

XSD2OWL and XML2RDF

ABox syntactic No No Semi-automatic Mapping rules

ConvertToRdf Population syntactic No No Semi-automatic Not mentioned

flat2rdf Population syntactic No No Semi-automatic Not mentioned

Java BibTeX-To-RDF Population syntactic No No Automatic Not mentionedGangemi et al. [GNV03, GGMO03]


DAML+OIL Single





RDF(S) / OWL Full

Single





RDF(S) Single



RDF(S) Single



OWL DL Single


[SLL 04, LS06] semantic ad-hoc relation







Java BibTeX To RDF Converter

Population syntactic No No Automatic Not mentioned

Excel2rdf Population syntactic No No Semi-automatic Not mentioned

RDF123 Population syntactic No No Semi-automatic Not mentioned

XLWrap Population syntactic No No Semi-automatic Mapping rules

19

Transformation











Several






• Ad-hoc solutions solutions for particular resources• Only focus on the resource type or

i l t ti

• No detailed guidelines for the selection of theresource are given

Th i f ti i tresource implementation •The resource provenance information is not considered

T f i O t lTransformationResearch Work



Single/ Several


Single



OWL DL Single



RDF(S) / OWL Full

Single


Ontology

• Do not include the three transformationapproaches (TBox, ABox, and Population).M k li it th hidd ti i th Gangemi et al. [GNV03,

GGMO03] Lightweight classes, attributes,

relations DAML+OIL Single





RDF(S) / OWL Full

Single





RDF(S) Single



RDF(S) Single



OWL DL Single


• Make explicit the hidden semantics in therelations of the NOR terms.

• Lack the application of re-engineeringpatterns

20











Several



Research Work Lightweight/ Heavyweight


Single/ Several

Hepp et al [HdB07] Lightweight classes relations RDF(S) / OWL Single

• ad-hoc solutions• only focus on the resource type or

i l t ti

• no detailed guidelines for the selection of theresourced t id th


Single



OWL DL Single

Abbasi et al. [ASC07] Lightweight instances Not mentioned SeveralMaala et al. [MDA07] Lightweight instances RDF SingleV A t l Li ht i ht l tt ib t RDF(S) / OWL Si l

Tool



Single/ Several

SKOS2GenTax Lightweight classes, attributes,

OWL DLP/ RDF(S)

Single

resource implementation • do not consider the resource provenance information



RDF(S) / OWL Full

Single



DAML+OIL Single



A l Li h i h l ib RDF(S) / OWL Si l

relationsT-ORG Lightweight instances Not mentioned SeveralKAON-REVERSE Lightweight classes,

attributes, relations,

F-Logic / RDF Single

T f i O t l



RDF(S) / OWL Full

Single





RDF(S) Single

l i h l ib (S) Si l

instancesODEMapster Lightweight instances RDF SingleD2R Server Lightweight instances RDF SingleTopBraid Composer Lightweight classes,

attributes, RDF/OWL (Full, DL or Lite)

Single

Transformation Ontology

• do not include the three transformationapproaches (TBox, ABox, and Population).

k li it th hidd ti i th



RDF(S) Single



OWL DL Single


Lightweight instances F-Logic / RDF Single

,relations, instances

)

XSD2OWL and XML2RDF

Lightweight classes, attributes, relations,


• make explicit the hidden semantics in therelations of the NOR terms.

• lack of the application of re-engineeringpatterns





An et al. [AM05] Lightweight instances RDF SingleCruz et al. [CXH04] Lightweight classes, attributes, RDF(S) Single

instancesConvertToRdf Lightweight instances RDF Singleflat2rdf Lightweight instances RDF SingleJava BibTeX-To- RDF Converter


21

relations, instances Foxvog et al. [FB06] Lightweight classes, attributes,

relations, instances CycL / OWL Full / WSML

Several

Excel2rdf Lightweight instances RDF SingleRDF123 Lightweight instances RDF SeveralXLWrap Lightweight instances RDF Several

Limitations of methods and toolsState of the Art


Ad-hoc solutions solutions for particular resourcesOnly focus on the resource type or

i l t ti

No detailed guidelines for the selection of theresourceare givenTh i f ti i tresource implementation The resource provenance information is not considered

T f i O t lTransformation Ontology

Do not include the three transformationapproaches (TBox, ABox, and Population).M k li it th hidd ti i th

Generate only ontology schema orontology instances

Make explicit the hidden semantics in therelations of the NOR terms.Lack the application of re-engineeringpatterns

22

Outline

23

Objectives and ContributionsObjectives & Research Methodology

ToTo speedspeed upup thethe ontologyontologydevelopmentdevelopment byby reusingreusing andand rere--engineeringengineering nonnon--ontologicalontologicalresourcesresourcesresourcesresources..Methodological objectives

O1. The definition of methodologicalaspects related to the reuse of non-ontological resources for buildingontological resources for buildingontologies.O2. The definition of methodologicalaspects related to the re-engineering ofnon ontological resources for buildingnon-ontological resources for buildingontologies.

Technological objectivesO3 The creation of a library of patterns forO3. The creation of a library of patterns forre-engineering non-ontological resourcesinto ontologies.O4. The development of a software library,NOR O that implements the suggestions

24

NOR2O, that implements the suggestionsgiven by the re-engineering patterns.

HypothesesObjectives & Research Methodology

H1 The reuse and re-engineering of NORs will allow the development of ontologies in an easierH1. The reuse and re engineering of NORs will allow the development of ontologies in an easier and faster way.

H2. It is possible to define a unified method for transforming NORs into ontologies i d d tlindependently

• (1) of the type, data model or implementation of the resource, and • (2) of the target ontology to be generated.

H3. The method for re-engineering NORs is extensible and adaptable to other types of resources.

H4. It is possible to create re-engineering patterns that allow generating ontologies fromH4. It is possible to create re engineering patterns that allow generating ontologies fromavailable NORs, namely, classification schemes, thesauri, and lexica, in an uniform way, independently

• (1) of how they have been implemented; • (2) of the target ontology to be generated• (2) of the target ontology to be generated,• (3) of the domain of the resources, that is, the patterns can be used to build ontologies in different

domains;

25

H5. The re-engineering patterns can be implemented in a software library that facilitates the work of ontology engineers when developing ontologies.

Research MethodologyObjectives & Research Methodology

Empirical InputsMethodological Inputs

Previous practices and experiences

g p

Available ad-hocmethods

Existing categorizationof resources

Empirical InputsMethodological Inputs

Previous practices and experiences

Available ad-hocmethods

Available Software Re-engineering practices

Methodological Inputs

26

g p

Available Software Re-engineering patterns

Previous practices in Ontology Engineering

Outline

27

Contributions

O1. The definition of methodological aspects related to the reuse of non-ontological resources for building ontologies.

28

Categorization of Non-ontological ResourcesContributions

• Non-ontological Resource (NOR) is a knowledge resource whose semantics has not yet been formalized by an ontology.

29

Non-ontological Resource Metadata Vocabulary (NoRMV)Contributions

OMV

30

http://www.oeg-upm.net/files/normv/normv.owl

Methodological guidelines for reusing NORsContributions

Non-ontological Resource Reuse

Non-ontological Resource Reuse refers to the processof choosing the most suitable non-ontological

Definition

resources for the development of ontologies.

To choose the most suitable non-ontological resourcesfor building ontologies.

Goal

Input OutputThe ontologyrequirementsspecification document(ORSD).

A set of non-ontologicalresources that to someextend covers theexpected domain.

WhoWho

Domain experts, software developers and ontologypractitioners.

When

After the ontology specification activity and before thenon-ontological resource re-engineering process.

31

Contributions

O2. The definition of methodological aspects related to the re-engineering ofnon-ontological resources for buildingnon-ontological resources for building ontologies.

32

Re-engineering Model for NORsContributions

Patterns for Re-engineeringNon ontological Resources

General Model for Software Re-engineeringNon-ontological Resources

(PR-NOR)

Ontology Forward

Engineering

Speci-

fication

Con-

ceptualNOR Reverse Engineering

Formalization

Conceptua-

lization

Design

TransformationRequirementsEngineering

ImplementationImplementation

Non-ontological Resource Ontology

RDF(S)

33

Template for the PR-NORContributions

Slot Value

General Information

Name Name of the pattern

Identifier An acronym composed of component type + abbreviated name of thecomponent + number

Component Type Pattern for Re-engineering Non-ontological Resource (PR-NOR)

Use Case

General Description in natural language of the re-engineering problem addressed by the pattern for re-engineering non-ontological resources.

Example Description in natural language of an example of the re-engineering problem.

Pattern for Re-engineering Non-ontological Resource INPUT: Resource to be Re-engineered

D i ti i t l l f th t l i lClassification scheme

General Description in natural language of the non-ontological resource.

Example Description in natural language of an example of the non-ontological resource.Graphical Representation

General Graphical representation of the non-ontological resource.

Example Graphical representation of the example of non-ontological resource.

OUTPUT: Designed Ontology

General Description in natural language of the ontology created after applying the

INPUTThesaurus

Lexicon

OUTPUTGeneralpattern for re-engineering the non-ontological resource.Graphical Representation

(UML) General Solution Ontology

Graphical representation, using the UML profile [BH06], of the ontology created for the non-ontological resource being re-engineered.

(UML) Example Solution Ontology

A graphical representation example, which uses the UML profile [BH06], of the ontology created for the non-ontological resource being used.

PROCESS: How to Re engineer

OUTPUT

Semantics of theRelations among the

NOR EntitiesPROCESS: How to Re-engineer

General Algorithm for the re-engineering process. Example Application of the algorithm to the non-ontological resource example.

Time Complexity The time complexity of the algorithm. Additional Notes Additional notes of the algorithm.

Formal Transformation

GeneralFormal description of the transformation made with the formal definitions of

PROCESSNOR Entities

Algorithm for the re-engineering process

34

General the resources.

Relationships (Optional)

Relations to other modelling

components Description of any relation to other PR-NOR patterns or other ontology design

patterns.

Semantics of the Relations among the NOR EntitiesContributions

TBox transformation: patterns must discover the semantics of the relations among the NOR entities.

1: Take two related terms from the NOR, ti and tj2: defaultRelation ← userDefinedRelation3: if contains(ti,tj) then4 l ti ti bCl Of tj

1.

5.

4: relation ← ti.subClassOf.tj5: else if contains(tj,ti) then6: relation ← tj.subClassOf.ti7: else

6.

8: wordnetRelation ← WordNet(ti, tj)9: if wordnetRelation == hyponym then

10: relation ← ti.subClassOf.tj11: else if wordnetRelation == hypernym then

1.

default relation

12: relation ← tj.subClassOf.ti13: else if wordnetRelation == meronym then14: relation ← ti.partOf.tj15: else if wordnetRelation == holonym then

3.

8.y16: relation ← tj.partOf.ti17: else18: relation ← defaultRelation19: end if 14

35

19: end if20: end if21: return relation

14.

Methodological guidelines for re-engineering NORsContributions

Non-Ontological Resource Re-engineering

Non-Ontological Resource Re-engineering refers tothe process of taking a non ontological resource and

Definition Cope with the classification schemes, thesauri, and lexica, in an uniform way, independently of how those resources have the process of taking a non-ontological resource and

transforming it into an ontology.

Creating an ontology from a non-ontological resource.

Goal

p ybeen implemented

Are based on re-engineering patterns

An ontology.

Input Output

One or more non-ontological resourcesselected by the reuseprocess.

Include the three transformation approaches (TBox, ABox and Population)

Make explicit the hidden semantics in the Who

Domain experts, software developers and ontologypractitioners.

When

prelations of the NOR terms by means of external resources in a semi-automatic way

Provide support for the transformation,

After the non-ontological resource reuse processand before the conceptualization activity.

When including information about who is in charge of performing a particular activity and when such an activity has to be carried out.

36

ContributionsMethodological guidelines for re-engineering NORs

Analyzing the non-ontological resource in order to identify its underlying components

37


The goal is to generate a conceptual model from the NOR

Criteria for searching for a suitable pattern:• NOR Type: classification scheme, thesauri, or lexicon• NOR Data model: C.Scheme data model, thesaurus data model,

38

or lexicon data model • Transformation approach: TBox, ABox


To output a new implementation of the ontology on the basis of the new conceptual model.

39

Contributions

O3. The creation of a library of patterns for re-engineering non-ontological resources into ontologies.

Patterns forclassification schemesn schemes

(8)

Patterns forPatterns forthesauri (4)

Patterns forlexica (4)

40

PR-NOR Library

Patterns for re-engineering classification schemesContributions

• A classification scheme1 is the descriptive information for an arrangement or division of objects into groups b d h t i ti hi h th bj t h i

Path Enumeration Data Model Adjacency List Data ModelN Identifier Type of NOR NOR Data

Model Target

1 PR-NOR-CLTX-01 Classification Path Ontology Schemabased on characteristics, which the objects have in common, e.g., water area classification scheme2.

Scheme Enumeration (TBox) 2 PR-NOR-CLTX-02 Classification

Scheme Adjacency List Ontology Schema

(TBox) 3 PR-NOR-CLTX-03 Classification Snowflake Ontology Schema

Snowflake Data Model Flattened Data Model

Scheme (TBox)4 PR-NOR-CLTX-04 Classification

Scheme Flattened Ontology Schema

(TBox) 5 PR-NOR-CLAX-10 Classification

S hPath Enumeration

Ontology (TB +AB )Flattened Data ModelScheme Enumeration (TBox+ABox)

6 PR-NOR-CLAX-11 Classification Scheme

Adjacency List Ontology (TBox+ABox)

7 PR-NOR-CLAX-12 Classification Scheme

Snowflake Ontology (TBox+ABox)

1 International Standard Organization (ISO) Information technology - Metadata registries – Part 1: Framework 2004

Scheme (TBox+ABox)8 PR-NOR-CLAX-13 Classification

Scheme Flattened Ontology

(TBox+ABox)

41

1. International Standard Organization (ISO). Information technology - Metadata registries – Part 1: Framework, 2004. Report ISO/IEC FDIS 11179-1.

2. http://www.fao.org/figis/servlet/RefServlet

Pattern for re-engineering a classification scheme followingthe adjacency list data model into an ontology schema

Contributions

ExamplePROCESS: How to Re-engineer

Require: Identification of the parent/child by using the adjacency list model 1: noParentTerms ← classification scheme terms without parent 2: if noParentTerms.length > 1 then 3: entityName ← name of the entity that contains the classification

h t

Require: Identification of the parent/child by using the adjacency list model 1: noParentT erms ← [Water area] 2: // noParentTerms.length=1 > 1 14: C 1 ← createClass(Water area) 16: children ← childrenOf(Water area) // using the adjacency list model 16: children ← [Environmental area; Jurisdiction area; Fishing statistical area]

scheme terms 4: rootClass ← createClass(entityName) 5: for ri ∈ noParentTerms do 6: Ri ← createClass(ri) 7: relation ← ExternalResource.getRelation(rootClass,Ri) 8: relate(relation,rootClass,Ri) 9: d f

[ ; ; g ]19: C 2 ← createClass(Environmental area) 21: rel1 ← ExternalResource.getRelation(C 1,C 2) 19: C 3 ← createClass(Jurisdiction area) 21: rel2 ← ExternalResource.getRelation(C 1,C 3) 19: C 4 ← createClass(Fishing statistical area) 21: rel3 ← ExternalResource.getRelation(C 1,C 4) 9: end for

10: end if 11: repeat 12: for cei ∈ noParentTerms do 13: if not alreadyCreatedClassFor(cei) then 14: C i ← createClass(cei) 15: end if

26: noParentTerms ← [Environmental area; Jurisdiction area; Fishing statistical area] 16: children ← childrenOf(Environmental area) // using the adjacency list model. 16: children ← [Inland/Marine;Ocean;North/South/Equatorial] 19: C 5 ← createClass(Inland/Marine) 21: rel4 ← ExternalResource.getRelation(C 2,C 5) 19: C 6 ← createClass(Ocean) 2115: end if

16: children ← childrenOf(cei) 17: for cej ∈ children do 18: if not alreadyCreatedClassFor(cej) then 19: C j ← createClass(cej) 20: end if 21: relation ← ExternalResource getRelation(cei cej)

21: rel5 ← ExternalResource.getRelation(C 2,C 6)19: C 7 ← createClass(North/South/Equatorial) 21: rel6 ← ExternalResource.getRelation(C 2,C 7) 16: children ← ∅ ← childrenOf(Jurisdiction area) 16: children ← childrenOf(Fishing statistical area) // using the adjacency list model 16: children ← [FAO Statistical area;Areal grid system] 19: C 8 ← createClass(FAO Statistical area)21: relation ← ExternalResource.getRelation(cei,cej)

22: relate(relation,cei,cej) 23: end for 24: add(allChildren,children) 25: end for 26: noParentT erms ← allChildren 27: removeAllTerms(allChildren)

19: C 8 ← createClass(FAO Statistical area)21: rel7 ← ExternalResource.getRelation(C 4,C 8) 19: C 9 ← createClass(Areal grid system) 21: rel8 ← ExternalResource.getRelation(C 4,C 9) 26: noParentTerms ← [Inland/Marine;Ocean;North/South/Equ.;FAO Statistical;Areal grid system] 16: children ← ∅ ← childrenOf(Inland/Marine)

42

27: removeAllTerms(allChildren)28: until isEmpty(noParentT erms)

( )16: children ← ∅ ← childrenOf(Ocean) 16: children ← ∅ ← childrenOf(North/South/Equatorial) 16: children ← ∅ ← childrenOf(FAO Statistical area) 16: children ← ∅ ← childrenOf(Areal grid system) 26: noParentTerms ← ∅

Patterns for re-engineering thesauriContributions

• Thesauri1 are controlled vocabularies of terms in a particular domain with hierarchical, associative and equivalence relations between terms. Thesauri

i l d f i d i d t i i f ti l i l d t b A

• Record-based data modelN Identifier Type of NOR NOR Data

Model Target

are mainly used for indexing and retrieving of articles in large databases. An example of thesaurus is the AGROVOC2 thesaurus.

9 PR-NOR-TSTX-01 Thesaurus Record-based Ontology Schema(TBox)

10 PR-NOR-TSTX-02 Thesaurus Relation-based Ontology Schema(TBox)(TBox)

11 PR-NOR-TSAX-10 Thesaurus Record-based Ontology (TBox+ABox)

12 PR-NOR-TSAX-11 Thesaurus Relation-based Ontology

• Relation-based data model

(TBox+ABox)

43

1. International Standard Organization (ISO). Documentation – Guidelines for the establishmentand development of monolingual thesaurus, 1986. Report ISO 2788.2. http://www.fao.org/agrovoc/

ContributionsPattern for re-engineering a thesaurus following

the relation-based data model into an ontology schemaPROCESS: How to Re-engineerExample

Require: Identification of the BT/NT/RT/UF relations by using the relation- based model

1: noBTerms ← terms without a broader term 2: repeat 3: for ti ∈ noBTerms do 4: if not alreadyCreatedClassFor(ti) then 5: C i ← createClass(ti)

Require: Identification of the BT/NT/RT/UF relations by using the relation- based model 1: noBTerms ← [Poaceae; Cereals] 5: C 1 ← createClass(Poaceae) 7: N T erms ← narrowerTermOf(Poaceae) 7: N T erms ← [Oryza] // using the relation-based model 12: C 11 ← createClass(Oryza) 14: rel1 ← ExternalResource.getRelation(C 1,C 11)5: C i ← createClass(ti)6: end if 7: NTerms ← narrowerTermOf(ti) 8: for tj ∈ NTerms do 9: if alreadyCreatedClassFor(tj) then

10: remove(N T erms,tj) 11: else 12 C j Cl ( j)

15: relate(rel1,C 1,C 11) 17: RTerms ← ∅ ← relatedTermOf(Poaceae) 27: UFTerms ← ∅ ← usedForTermOf(Poaceae) 31: restOfTerms ← [Oryza] 5: C 2 ← createClass(Cereals) 7: NTerms ← narrowerTermOf(Cereals) 7: NTerms ← [Rice] // using the relation-based model 12: C 21 ← createClass(Rice)12: C j ← createClass(tj)

13: end if 14: relation ← ExternalResource.getRelation(C i,C j) 15: relate(relation,C i,C j) 16: end for 17: RTerms ← relatedTermOf(ti) 18: for tr ∈ RTerms do

12: C 21 ← createClass(Rice)14: rel2 ← ExternalResource.getRelation(C 2,C 21) 15: relate(rel2,C 2,C 21) 17: RTerms ← ∅ ← relatedTermOf(Cereals) 27: UFTerms ← ∅ ← usedForTermOf(Cereals) 31: restOf T erms ← [Oryza;Rice] 34: noBTerms ← restOfTerms ← [Oryza;Rice] 35: removeAllTerms(restOfTerms)

19: if alreadyCreatedClassFor(tr) then20: remove(RTerms,tr) 21: else 22: C r ← createClass(tr) 23: end if 24: relation ← ExternalResource.getRelation(C i,C r) 25: relate(relation,Ci,Cr)

( )4: // Oryza class, C 11, already created 7: NTerms ← ∅ ← narrowerTermOf(Oryza) 17: RTerms ← relatedTermOf(Oryza) 17: RTerms ← [Rice] // using the relation-based model 20: remove(RTerms,Rice) // Rice class, C21, already created 24: rel3 ← ExternalResource.getRelation(C 11,C 21) 25: relate(rel3,C 11,C 21) 27 UFT ∅ dF T Of(O )

( )26: end for 27: UFTerms ← usedForTermOf(ti) 28: for tq ∈ UFTerms do 29: SOE(ti,tq) 30: end for 31: add(restOfTerms,NTerms) 32: add(restOfTerms,RTerms)

27: UFT erms ← ∅ ← usedForTermOf(Oryza)31: restOfTerms ← ∅ 4: // Rice, C 21, already created 7: NTerms ← ∅ ← narrowerTermOf(Rice) 17: RTerms ← ∅ ← relatedTermOf(Rice) 27: UFTerms ← usedForTermOf(Rice) 27: UFTerms ← [Paddy] // using the relation-based model 29: SOE(Rice Paddy)

44

32: add(restOfTerms,RTerms)33: end for 34: noBTerms ← restOfTerms 35: removeAllTerms(restOfTerms) 36: until isEmpty(noBTerms)

29: SOE(Rice,Paddy)31: restOfTerms ← ∅ 34: noBTerms ← ∅ ← restOfTerms 35: removeAllTerms(restOfTerms)

Patterns for re-engineering lexicaContributions

• A lexicon1 is a list of words in a language (a vocabulary) along with some knowledge of how to use each word. A lexicon may be general or domain-specific; we might have, for example, a lexicon of several thousand common

• Record-based data modelN Identifier Type of NOR NOR Data

ModelTarget

words of English or German, or a lexicon of the technical terms of dentistry in some language. An example is WordNet2

• Relation-based data model

Model 13 PR-NOR-LXTX-01 Lexicon Record-based Ontology Schema

(TBox) 14 PR-NOR-LXTX-02 Lexicon Relation-based Ontology Schema

(TBox)(TBox)15 PR-NOR-LXAX-10 Lexicon Record-based Ontology

(TBox+ABox) 16 PR-NOR-LXAX-11 Lexicon Relation-based Ontology

(TBox+ABox)(TBox+ABox)

45

1. G. Hirst. Ontology and the lexicon. In Handbook on Ontologies in Information Systems, pages 209–230. Springer, 2004.2. http://www.wordnet-online.com

Pattern for re-engineering a lexicon following the record-based data model into an ontology schema

Contributions

PROCESS: How to Re-engineer

Require: Identification of the relations by using the record-based model 1: Synsets ← all the synsets of the lexicon 2: for si ∈ Synsets do 3: if not alreadyCreatedClassFor(si) then C i ← createClass(si) endif 4: Hyponyms ← hyponymOf(si)4: Hyponyms ← hyponymOf(si)5: for sj ∈ H yponyms do 6: if not alreadyCreatedClassFor(sj) then C j ← createClass(sj) endif 7: relate(subClassOf,C i,C j) 8: end for 9: Hypernyms ← hypernymOf(si)

10: for sk ∈ Hypernyms do 11 if t l d C t dCl F ( k) th C k ← t Cl ( k) dif

Example

Require: Identification of the relations by using the record-based model 1: Synsets ← [cell;cell part;animal cell] 3: C 1 ← createClass(cell)11: if not alreadyCreatedClassFor(sk) then C k ← createClass(sk) endif

12: relate(subClassOf,C k,C i) 13: end for 14: Meronyms ← meronymOf(si) 15: for sl ∈ Meronyms do 16: if not alreadyCreatedClassFor(sl) then C l ← createClass(sl) endif 17: relate(partOf,C i,C l)

( )4: Hyponyms ← hyponymOf(cell) 4: Hyponyms ← [animal cell] // using the record-based model 6: C 2 ← createClass(animal cell) 7: relate(subClassOf,C 2,C 1) 9: Hypernyms ← ∅ ←hypernymOf(cell) 14: Meronyms ← meronymOf(cell) // using the record-based model

18: end for19: Holonyms ← holonymOf(si) 20: for sm ∈ Holonyms do 21: if not alreadyCreatedClassFor(sm) then C m ← createClass(sm) enif 22: relate(partOf,C m,C i) 23: end for 24: Synonyms ← synonymOf(si)

14: Meronyms meronymOf(cell) // using the record based model14: Meronyms ← [cell part] 16: C 3 ← createClass(cell part) 17: relate(partOf,C 3,C 1) 19: Holonyms ← ∅ ← holonymOf(cell) 24: Synonyms ← synonymOf(cell) // using the record-based model 24: Synonyms ← [CESP:C16C10 8]: Sy o y s sy o y O (s )

25: for sn ∈ Synonyms do 26: CorchoEtAlPattern(si,tn) // Corcho et al. [Cor09]. Logical Pattern 27: end for 28: RelatedSynsets ← relatedSynsetOf(si) 29: for so ∈ RelatedSynsets do 30: if not alreadyCreatedClassFor(so) then C o ← createClass(so) endif 31: relation ← ExternalResource getRelation(C i Co)

24: Synonyms ← [CESP:C16C10.8]26: CorchoEtAlPattern(cell,CESP:C16C10.8) // Corcho et al. [Cor09]. 28: RelatedSynsets ← ∅ ← relatedSynsetOf(cell)

46

31: relation ← ExternalResource.getRelation(C i,Co)32: relate(relation,C i,C o) 33: end for 34: end for

PR-NOR library at the ODP PortalContributions

47

http://ontologydesignpatterns.org/wiki/Submissions:ReengineeringODPs

Contributions

O4. The development of a software library, NOR2O, that implements the suggestions given by the re-suggestions given by the re-engineering patterns.

48

NOR2O Software LibraryContributions

49

Outline

50

Controlled experiment in Master courseEvaluation

Controlled experiment in academic setting to validate Participants create an ontology schema from theControlled experiment in academic setting to validate • understandability• applicability • usability

Participants create an ontology schema from the ETT thesaurus, which follows the record-based data model and is implemented in XML.

H1H2

Setting Participants from Master Course. Build manually a conceptual model from a thesaurus. Analysing the guidelines and patterns.

Results Conceptual model from an excerpt of a thesaurus in

H2

razonable time, 30 min.

Conclusions The methodological guidelines seem to be useful and understandable.

Examplep

Require: Identification of the BT/NT/RT/UF relations by using the relation- based model 1: noBTerms ← [Poaceae; Cereals] 5: C 1 ← createClass(Poaceae) 7: N T erms ← narrowerTermOf(Poaceae) 7: N T erms ← [Oryza] // using the relation-based model 12: C 11 ← createClass(Oryza) 14: rel1 ← ExternalResource.getRelation(C 1,C 11) 15: relate(rel1,C 1,C 11) 17: RTerms ← ∅ ← relatedTermOf(Poaceae) 27: UFTerms ← ∅ ← usedForTermOf(Poaceae) 31: restOfTerms ← [Oryza] 5: C 2 ← createClass(Cereals) 7: NTerms ← narrowerTermOf(Cereals) 7: NTerms ← [Rice] // using the relation-based model 12: C 21 ← createClass(Rice) 14

H4

14: rel2 ← ExternalResource.getRelation(C 2,C 21)15: relate(rel2,C 2,C 21) 17: RTerms ← ∅ ← relatedTermOf(Cereals) 27: UFTerms ← ∅ ← usedForTermOf(Cereals) 31: restOf T erms ← [Oryza;Rice] 34: noBTerms ← restOfTerms ← [Oryza;Rice] 35: removeAllTerms(restOfTerms) 4: // Oryza class, C 11, already created 7: NTerms ← ∅ ← narrowerTermOf(Oryza) 17: RTerms ← relatedTermOf(Oryza) 17: RTerms ← [Rice] // using the relation-based model 20: remove(RTerms,Rice) // Rice class, C21, already created 24: rel3 ← ExternalResource.getRelation(C 11,C 21) 25: relate(rel3,C 11,C 21) 27: UFT erms ← ∅ ← usedForTermOf(Oryza) 31: restOfTerms ← ∅ 4: // Rice C 21 already created

PR-NOR-TSTX-01

51

4: // Rice, C 21, already created 7: NTerms ← ∅ ← narrowerTermOf(Rice) 17: RTerms ← ∅ ← relatedTermOf(Rice) 27: UFTerms ← usedForTermOf(Rice) 27: UFTerms ← [Paddy] // using the relation-based model 29: SOE(Rice,Paddy) 31: restOfTerms ← ∅ 34: noBTerms ← ∅ ← restOfTerms 35: removeAllTerms(restOfTerms)

SEEMP ProjectEvaluation

Build a Reference Ontology by applying our method for reusing and re-engineering NORs

H1H2H2

Ontology Concepts Attributes Axioms Instances Efforts(man.months)

SEEMP RO 1985 315 1037 1449 6

Resource Type Data Model Implementation Pattern usedNACE Classification Path enumeration Database PR-NOR-CLTX-01

SchemeFOET Classification

Scheme Path enumeration Database PR-NOR-CLTX-01

ISCED 97 ClassificationScheme

Adjacency list Database PR-NOR-CLTX-02

ISO 4217 ClassificationScheme

Snowflake XML PR-NOR-CLAX-12

H4





CEFR ClassificationScheme

Proprietary model Proprietary format

EU Driving Classification Snowflake Proprietary for-

52

gLicense Scheme

p ymat

EURES Skill ClassificationScheme

Path enumeration Database PR-NOR-CLTX-01

LE FOREMContracts

Proprietaryclassification

Proprietary model Proprietary format

GeoLinkedData ProjectEvaluation

Populate the SCOVO vocabulary from INE datasetsPopulate the SCOVO vocabulary from INE datasets

H1H2

H3

H5

53

H5

Controlled experiment to validate the quality of the patterns and NOR2OEvaluation

Setting Twoontologyengineers.Setting Two ontology engineers. Build five ontologies from available NORs. Compare these five gold standard ontologies against the ontologies generated by NOR2O.

Results Similarity values of every ontology generated with the goldstandardontology

Ontology Engineer

Ontology Engineersgold standard ontology.

Conclusions The ontologies generated have an acceptable level of quality, meaning by quality the similarity of the ontologies to the gold standard ones.

Name Type Data Model Implementation N of N of Pattern

NOR2O

Name Type Data Model Implementation N. of terms

N. of terms covered

Pattern Used

ASFA thesaurus record-based XML 9882 188 PR-NOR-TSTX-01 ETT thesaurus record-based XML 2522 337 PR-NOR-TSTX-01 ACM classification adjacency list XML 1606 223 PR-NOR-CLTX-02ACM classification

scheme adjacency list XML 1606 223 PR-NOR-CLTX-02

FOET classification scheme

path enumeration spreadsheet 127 112 PR-NOR-CLTX-01

BioLexicon lexicon relation-based database 53876 150 PR-NOR-LXTX-02 H4

Similarity values between ontologies generated with the gold standard

Cider StrucSubsDistAlignmentASFA 0.754 0.631ETT 0.713 0.745

similarity

H5

54

ACM 0.620 0.870FOET 0.621 0.753

BioLexicon 0.515 0.793

Outline

55

Comparative analysis of the three most representative methods and oursConclusions

Features Heep et al. Hyvo nen et al. Soerger et al. Villazo n- Terrazas

Transformation Transformation TBox TBox TBox TBox, ABox,

approach , ,

Population Transformation

aspects syntactic semantic

syntacti semantic

syntactic semantic

syntactic, semantic

Semantics of subClassOf, subClassOf, subClassOf, subClassOf, the NOR relations

ad-hoc relation partOf ad-hoc relation partOf

Additional resources/Do-

No DOLCE Domain expert WordNet

main expert

Automatic / Semiautomatic

/ Manual

Semiautomatic Semiautomatic Manual Semiautomatic

/ Manual Technique Not mentioned Not mentioned Not mentioned Re-engineering

patterns Tool support SKOS2GenTax ad-hoc tool Not mentioned NOR2 O

56

Outline

57

Creation of a registry of non ontological resources

Future work

Creation of a registry of non-ontological resources.

Re-engineering NORsg gRicher ontologies

Additional external resources e g DBpediaAdditional external resources, e.g., DBpedia

Generation of GoodRelations-compliant ontologies

Multilingual NORs

Integration of different NORsIntegration of different NORs

Evolution of the NORs

58

Linked Data aspects

Outline

59

Selected Publications & Patent1 B. Villazón-Terrazas J Ramírez M C Suárez-Figueroa A Gómez-Pérez (2011) “A Network of1 B. Villazón Terrazas, J. Ramírez, M.C. Suárez Figueroa, A. Gómez Pérez (2011). A Network of

Ontology Networks for building e-Employment Advanced Systems”. In International Journal onExpert Systems with Applications, ELSEIVER.

2 B. Villazón-Terrazas, M.C. Suárez-Figueroa, A. Gómez-Pérez (2010). “A Pattern-Based Method for Re-engineering Non-ontological Resources into Ontologies”. In International Journal on Semantic g g g gWeb and Information Systems, Wright State University, USA.

3 B. Villazón-Terrazas, A. Gómez-Pérez, Jean Paul Calbimonte (2010). “NOR2O: a Library for Transforming Non-ontological Resources to Ontologies”. Poster In Extended Semantic Web Conference (ESWC2010)

4 L. M. Vilches Blázquez, B. Villazón-Terrazas, V. Saquicela, A. de León, O. Corcho, A. Gómez-Pérez (2010) . “GeoLinked data and INSPIRE through an application case”. In ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems (GIS2010)

5 B. Villazón-Terrazas, M.C. Suárez-Figueroa, A. Gómez-Pérez (2009). “Pattern for Re-engineering a Term-based Thesaurus, which follows the Record-based model, to a Lightweight Ontology”. In Workshop on Ontology Patterns (WOP). In International Semantic Web Conference (ISWC 2009)

6 A. Gómez-Pérez, J. Ramírez, B. Villazón-Terrazas (2007). “An Ontology for Modelling Human Resources Management based on Standards”. In 11th International Conference on Knowledge-Based Intelligent Information & Engineering Systems (KES2007)Based Intelligent Information & Engineering Systems (KES2007)

7 A. Gómez-Pérez, J. Ramírez, B. Villazón-Terrazas (2007) “Methodology for Reusing HumanResources Management Standards”. In: 19th International Conference on Software Engineeringand Knowledge Engineering (SEKE 2007)

8 B. Villazón-Terrazas, A. Gómez-Pérez, J. Ramírez (2009) “Ontología de CV y Ofertas de Empleo”.

60

8 B. Villazón Terrazas, A. Gómez Pérez, J. Ramírez (2009) Ontología de CV y Ofertas de Empleo . Patent, Registro de la Propiedad Intelectual. Comunidad de Madrid. Reference M-404/2009.

A Method for Reusing and Re-engineering Non-ontological

Resources for Building Ontologies

Boris Villazón-Terrazas

[email protected]

th14th April, 2011

a method for reusing and re-engineering non-ontological resources for building ontologies

Technology