Ontology Engineering Resource Re-Engineering

Ontology Engineering and
Resource Re-Engineering
Asunción Gómez-Pérez (asun@fi.upm.es)
Credits to:
Boris Villazón-Terrazas (bvillazon@fi.upm.es)
Mari Carmen Suárez -Figueroa (mcsuarez@fi.upm.es)

Work distributed under the license Creative Commons Attribution-Noncommercial-Share Alike 3.0
Ontological Engineering 1 © A. Gómez Pérez, M.C. Suarez de Figueroa, B. Villazón, E.Montiel, G. Aguado, M. Espinoza

 Motivation Index
 State of the Art on Methodologies for Ontology Engineering
 Building an ontology
 Conclusion


Motivation

Classical
In our team, we want to build an OWL ontology in
the geographical domain, but we want to use
several geographical standardized terminology
stored in DBs and classification schemes in our
own format. I want to combine it with statistic
information about unemployment

Non Ontological
Resource Reuse

Non Ontological
Resource
Reengineering


Where is the terminology?
DBs from a Geographical institute

Catalog/ID

Implicit knowledge
Heterogeneity coded in numbers
.- Language
.- Terminology XX-YY-ZZ
.- Granularity 02-01-02
02: transportation
01: road
02: 3-lines highway


Different Data Models
with the same information


Lot of terminologies being used …

Music On-line activities

E-Gov Publications
Cross-domains

Geographic Life Sciences


Tools for searching vocabularies

http://schemaweb.info

http://swoogle.umbc.edu/

Linked Open Vocabularies

http://labs.mondeca.com/dataset/lov

http://watson.kmi.open.ac.uk/WatsonWUI/

7

After some proces….


Geolinkeddata ontologies

Ontology Network
WGS84 Geo
Positioning: an RDF
vocabulary scv:Dimension
scv:Item
W3C 4 scv:Dataset

hydrographical Vocabulary
Time ontology phenomena (rivers,
lakes, etc.)
O.
hasStatisticalData Statistics
hasLat/Long WGS84 hasLat/Long
SCOVO
SCOVO
on
hydrOntology hasLocation/isLocated FAO
Vocabulary for
hydrOntology O. instants, intervals,
durations, etc.
FAO Time
UNESCO Geopolitical
EGM / ERM 4 hasGeometry
hasGeometry
ontology
W3C Time
GeoNames
… GML
Names and Legend
international code
Ontology for OGC GML 4 systems for
territories and
Ontology
Geography Markup Specification
Language groups
Specification 4
Thesaurus


Motivation
Ontology
Mappings

In our team, we want to build an ontology about the
Classical
human resources management domain. The
ontology should include information about
occupations and activity sectors, data must be kept
in the original DBs, and we want to have the
ontology in several natural languages.

Ontology-DB
mapping Ontological
Resource Reuse

Re-engineering
Non-ontological Localizing Ontological
resources Resources


Index

 Motivation
 Conclusion


Building ontologies in the 1990s and 2000s

Methodologies for building single ontologies do not
consider the reuse of knowledge
• Uschold and King’s method
• Grüninger and Fox’s methodology
• KACTUS approach
• METHONTOLOGY
• SENSUS method
• On-To-Knowledge
• DILIGENT

Ontology learning approaches for building ontologies from
structured, semi-structured and non-structured data
• Are not integrated with current methodologies
• Mainly from non-structured data using NLP techniques

Current situation
• Reuse of knowledge-aware resources
– Already consensuated by domain experts
• Ontologies are built collaboratively
• Ontologies are connected in ontology
networks
• Multilingual features

Ontology
Development
Process


Knowledge Resources
NeOn Scenarios
Non Ontological Resources Ontological Resources

Glossaries Dictionaries Lexicons O. Design Patterns O. Repositories and Registries
3 4
Flogic
5 6
Classification
Schemas
Taxonomies Thesauri RDF(S)
OWL
Ontological Resource
2 Reuse

2 5 6
O. Aligning
Non Ontological Resource Ontology Design 4
Reuse Pattern Reuse 3
6 O. Merging

2 Ontological Resource
7 Reengineering
5 Alignments
Non Ontological Resource
Reengineering 4 6

1
RDF(S)
O. Specification Scheduling O. Conceptualization O. Formalization O. Implementation
Flogic
8
9 Ontology Restructuring
(Pruning, Extension,
O. Localization Specialization, Modularization) OWL

1,2,3,4,5,6,7,8, 9
14
Ontology Support Activities: Knowledge Acquisition (Elicitation); Documentation;
OntologicalConfiguration Management; Evaluation
Engineering 14 (V&V); Assessment Suarez de Figueroa, B. Villazón, E.Montiel, G. Aguado, M. Espinoza
© A. Gómez Pérez, M.C.

NeOn Methodology

Process and activities covered: All processes and activities are described with:
 Ontology Specification  A filling card
 Scheduling  A workflow
 Non-Ontological Resource Reuse  Examples
 Non-Ontological Resource Re-engineering
 Reuse General Ontologies
 Reuse Domain Ontologies
 Reuse Ontology Statements
 Reuse Ontology Design Patterns


Handbook on the NeOn Methodology

http://www.neon-project.org/nw/NeOn_Book
Ontological Engineering 16
Slide© A. Gómez Pérez, M.C. Suarez de Figueroa, B. Villazón, E.Montiel, G. Aguado, M. Espinoza
16

 22 Executive Chapter Summaries are available at the NeOn Web Site

Slide 17
Slide 17

Index

 Motivation
 Conclusion


Looking for an European Employment
o
EURES
p ESES
(Int) n Cand.
ES ES
m
q
ES  Cand. ES
Vacan.

Vacan. Cand. l
Catalonia
r ESES
(Es)
ES Vacan.

i
Lombard ES
ES (It)
Cand.

a h
ES ES

LEGENDA
b
ES

g
Private
ES ES
(Int) 
ES Employment Service c f Vacan.
Cand.
Cand. ES ES
Job Seeker’s Candidacy d e
Wallonia
ES ESES
(Be)
Vacan.
Vacan.
Employer Job Vacancy

Helping Job Seekers on their way
Local
o
EURES Matching
p ESES n
(Int) algorithm
ES ES
m Local
Matching
q ES algorithm
Cand. ES Cand.
Cand.
Vacan. l
Catalonia
r Vacan.
Vacan.
Cand.
Cand. ESES
(Es)
ES

Cand.
Cand.
European Vacan.
Vacan.
i
Lombard Employment ES
ES (It) Mediators
Vacan.
Vacan.
Marketplace Cand.
Cand.

a h
ES ES
LEGENDA Cand.
Cand.
Vacan.
Vacan.
Vacan.
Vacan. Local
Matching
Requester ES b Private algorithm
g
ES ES ES
(Int)
Responding ES Local
Matching
ES not involved c f
algorithm
Cand.
Cand. ES ES
d e
Wallonia
Job Seeker’s Candidacy ESES
(Be)
ES
Vacan.
Vacan.
Employer Job Vacancy

Key issues

• Reuse of proprietary knowledge-aware
resources
• Heterogeneity
– Terms are in different languages
– Different conceptualization (different ways of
organizing job categories)
– Different DB schemas
• Data must be kept in the original sources
and in their own language.


Key aspects of Ontological Engineering
• Ontologies
– Single versus network of ontologies?
– Are ontologies built from scratch or reusing knowledge-aware
resources?
– Are mappings used for solving conceptual mistmaches?
• Instances
– Where are the data/instances?
• Instances are in the ontology
• Instances are in RDF files independently of the ontology
• Data are kept in the original sources
– Are instances distributed or centralized?
– Have instances a very high rate of changes?
– Heterogeneous provenance of instances
– Degrees of data quality
– Permissions


Centralized network of Federated network of ontologies
ontologies where data are where data are distributed
distributed
Ms Ms

Ms
Ms Ms
Ms

Ms
Ms Ms
Ms
Ms

Ms

Ms
Ms Ms

1. Build a reference ontology for the domain
1. Build a reference ontology 2. Build local ontologies
3. Build mappings between the core and local ontologies
2. Build mappings between the 4. Build mappings between the local ontologies and the data
reference ontology and the data sources sources


Knowledge Resources
NeOn Methodology

3 4
Flogic
5 6
Classification
Schemas
OWL
2 Reuse

2 5 6
O. Aligning
6 O. Merging

7 Reengineering
5 Alignments
Reengineering 4 6

1
RDF(S)
Flogic
8

1,2,3,4,5,6,7,8, 9
24

Ontology Requirement
Specification Document
Ontology Requirements Specification Document Template

1 Purpose
“Software developers and ontology practitioners should include in
this slot the purpose of the ontology”
2 Scope
this slot the scope of the ontology”
3 Level of Formality
this slot the level of formality of the ontology”
4 Intended Users
this slot the intended users of the ontology”
5 Intended Uses
this slot the intended uses of the ontology”
6 Groups of Competency Questions
this slot the groups of competency questions and their answers,
including priorities for each group”
7 Pre-Glossary of Terms
Terms
this slot the list of terms included in the CQs and their
frequencies”

Ontology Specification:
Identify requirements using competency questions


Ontology Specification.
The Ontology Requirement Specification
Document


Knowledge Resources
NeOn Methodology

3 4
Flogic
5 6
Classification
Schemas
OWL
2 Reuse

2 5 6
O. Aligning
6 O. Merging

7 Reengineering
5 Alignments
Reengineering 4 6

1
RDF(S)
Flogic
8

1,2,3,4,5,6,7,8, 9
28

Searching Resources
• Use the terminology from the ORSD

• Find resources covering the terminology
Knowledge Resources


Flogic
Classification
Schemas
OWL
2

• Where:
- Internet
- Standardization bodies (ISO,…)
- Intranet of the organization
- Ontology Registries

Search and Select
non-ontological resources

• We select the most appropriate
standards and taxonomies for:
– Occupation Classification Language Classification
ISCO-88 (COM), SOC, ISCO-88, ISO 6392, CEF
ONET, Eures Taxonomy.
– Classification of Economic
Driving License Classification
Activities European Legislation
ISIC Rev. 3.1, NACE Rev. 1.1, Skill Classification
NAICS
– Apprenticeship classifications
Eures Taxonomy
ISCED 97, FOET Contract Types Classification
– Currency Classification LE FOREM, Eures and BLL Classification
ISO 4217 Work Condition Classification
– Geography Classification LE FOREM, Eures and BLL Classification
ISO 3166, Eures Taxonomy

Is the terminology included in
the Ontology Requirements Specification Document
covered by the resources?

ISO 4217 (currencies) ISO 3166 (countries)


Multilingual Non-ontological resources - ISCO-88 (COM)


Searching Ontologies
Ontology Requirement
Specification Document

• Search ontologies
• Compare and Assess ontologies
• Select the best ontology based on
– Coverage of the domain
– Expressivity of the Implementation language

Knowledge Resources
NeOn Methodology

3 4
Flogic
5 6
Classification
Schemas
OWL
2 Reuse

2 5 6
O. Aligning
6 O. Merging

7 Reengineering
5 Alignments
Reengineering 4 6

1
RDF(S)
Flogic
8

1,2,3,4,5,6,7,8, 9
34

Gantt chart for your project. Waterfall model


Reuse and Re-engineer ing + Incr emental

Ontological Engineering 36 Slide 36
© A. Gómez Pérez, M.C. Suarez de Figueroa, B. Villazón, E.Montiel, G. Aguado, M. Espinoza

Knowledge Resources
NeOn Methodology

3 4
Flogic
5 6
Classification
Schemas
OWL
2 Reuse

2 5 6
O. Aligning
6 O. Merging

7 Reengineering
5 Alignments
Reengineering 4 6

1
RDF(S)
Flogic
8

1,2,3,4,5,6,7,8, 9
37

Motivation

I want to transform my
adjacency list-based
classification into an
ontology


Types of non-ontological resources

Non-Ontological
Resources are
knowledge-aware
resources whose
semantics have not
been formalized yet
by means of an
ontology


Types of non-ontological resources


Approaches to transform
resources into ontologies

ABox

TBox

Population


Approach for Re-engineering Non-Ontological
Resources


Patterns for Re-engineering Classification Schemes into Ontologies

– ABox transformation

– TBox transformation


Re-engineering NORs
Template for the PR-NOR

INPUT

OUTPUT

PROCESS

© A Method for Reusing and Re-engineering Non-Ontological

Patterns for Re-engineering Classification Schemes
Pattern for re-engineering a classification scheme, which follows the adjacency list data model,
into an ontology schema

INPUT: Non-Ontological Resource
OUTPUT: Ontology PROCESS: How to Re-
General Generated
General engineer
Example

Example
Example


Technological support
PR-NOR library at the ODP Portal

NOR2O: a Library for Transforming
Non-Ontological Resources to Ontologies
http://mccarthy.dia.fi.upm.es/nor2o/

Pattern based approach for re-engineering
non ontological resources
ISCO-88 (COM)
Inter national Standar d Classification
of Occupations Occupation
Ontology
(for Eur opean Union pur poses)

FOET
Classification of fields of Education
Ontology
education and tr aining
Pattern for re-engineering a
NACE classification scheme modelled
Statistical Classification of with a Path Enumeration Data Model
Economic Activities in the Economic
Eur opean Community Activity
Ontology

ISO 3166 Pattern for re-engineering a
English countr y names
classification scheme modelled
and code elements Geography
with a Snowflake Data Model Ontology

ISTAT
Italian Geogr aphy
Standar d
Pattern for re-engineering a
classification scheme modelled Italian
Geography
with an Adjacency List Data Model Ontology


Knowledge Resource Re-engineering and
Aggregation

…. ISO 3166-1 (XML) Regions Table
<ISO_3166-1_Entry>
(Eures Oracle DB)
<ISO_3166-1_Country_name>SPAIN</ISO_3166-1_Country_name>
<ISO_3166-1_Alpha-2_Code_element>ES</ISO_3166-1_Alpha-2_Code_element>
</ISO_3166-1_Entry>
…

Excerpt of the
Location
Geography Ontology
Ontology model subClass-Of

Country has region Region

Spain Cataluña
Canarias
Ontology instances
Galicia
Andalucía

Knowledge Resources
NeOn Methodology

3 4
Flogic
5 6
Classification
Schemas
OWL
2 Reuse

2 5 6
O. Aligning
6 O. Merging

7 Reengineering
5 Alignments
Reengineering 4 6

1
RDF(S)
Flogic
8

1,2,3,4,5,6,7,8, 9
49

Conceptualization:
Modular approach for ontology construction

Reusability Usability

- +
Application
Domain O. : Job Seeker, Job Offer

Domain O.: Economic Activity, Occupation, Education, Skill, Driving
License, Compensation, Labour Regulatory, Competence

General/Common Ontologies: Time, Geography, Language

+ Representation Ontology: WSML -


Reference Ontology ISCO-88 COM
EURES CEF
ONET
ISO 6392
EURES

Language
Skill
Ontology
Ontology Occupation
Ontology
EURES
LE FOREM + BLL + EURES
ISO 3166

subClass-Of

Labour
Regulatory Geography
Ontology Ontology

DAML Time
Ontology

Job Offer Job Seeker
Ontology Ontology
Competence has date of birth
Ontology / is date of birth of

Time
Ontology
ISO 4217

FOET
subClass-Of
ISCED97
Compensation
Ontology

European Legislation
Education
Ontology
Driving Ad hoc wrapper
License Economic
Ontology External Sources
Activity
Ontology NACE Rev. 1.1


Details of the ontology
has education / requires education /
is education of Education is associated with
Language has competence / requires competence /
Ontology is competence of Competence is associated with
Competence
has mother language / Ontology
Education

has work condition /
Language
is mother tongue of Ontology has work condition /

is associated to
Work Condition Labour is associated with
Regulatory

has contract type /
Job Offer

has compensation / is associated to
Ontology Ontology
speaks /
is spoken by Job Seeker has contract type /
Contract Type
Ontology is associated with
has compensation /
is associated with
has nationality from / Compensation has location /
Job Seeker
is nation of Compensation is location of
Organization

is associated with / is associated to
Ontology
has candidacy/ resides in / has job vacancy/
belongs to is residence of Country belongs to
Geography
Ontology
Candidacy
has work experience /
Job Vacancy
belongs to
Location

has location
is located in /
has objective / has activity sector / is associated with has vacancy/
has job category / belongs to is associated with belongs to
Offered Work Requested
Sector
is associated with Experience Work Experience

Objective is associated with / Vacancy
has activity sector / requires work experience
is associated with
has activity sector /
subClass-Of is associated with subClass-Of
Economic
has activity sector /
Activity
is associated with
has job category Computing Ontology has job category/ ICT
ICT Objective
is associated with Professionals is associated with Vacancy

has job category / subClass-Of Occupation has job category/
Ontology is associated with
is associated with
Occupation
has job category /
is associated with

Conclusions

1. The NeOn methodology
facilitates the reuse and
reengineering of non
ontological resources into
ontologies

2. The reuse of non-ontological
resources that have been
reached some degree of
consensus in a community
allows the development of
ontologies easier and quicker


Ontology Engineering Resource Re-Engineering

Recomendados

Recomendados

Mais conteúdo relacionado

Último

Último (20)

Destaque

Destaque (20)

Ontology Engineering Resource Re-Engineering