INSPIRE 2012-Gülten Kara

7/29/2019 INSPIRE 2012-Glten Kara

1/35

SEMANTIC DEFINITION AND MATCHING

FOR NATIONAL SPATIAL DATA

INFRASTRUCTURE

Glten KARA1, Deniztan ULUTA2,etin CMERT1

[email protected], [email protected], [email protected]

1

Karadeniz Technical University, Engineering Faculty, Geomatics Engineering, Trabzon, TURKEY2Computer Science Department, Semantic Web Lab. University of Texas at Dallas, USA

http://www.harita.ktu.edu.tr/
mailto:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]


2/35

Outline

Introduction Ontology Why Semantic? Semantic Definition Projects and Studies

Semantic Web Requirements Methodology for Semantic Definition Semantic Web Languages Implementation Architecture

Transformation Problems-Tools Matching Results Future Works and Conclusion


3/35

Introduction

In Turkey, the establishment of National Spatial Data Infrastructure (NSDI)

is on the agenda.

The technologies used for technological infrastructure of any SDI areSyntactic Webtechnologies.

In the near feature, the current technologies will be replaced by SemanticWeb technologies.


4/35

Ontology

An ontology represents a set of concepts within a domain, the

relationships between these concepts and the constraints on theproperties. In accordance with RDF or OWL, ontologies are made uptwo main components. Ontologies can assist in communication betweenpeople and computers.

LineRoad

Classes Relationships Classes

hasGeometry


5/35

Why Semantic?

In the Information Age, the importance given to knowledge

is rapidly increasing. But, information sources are vastlyvaried and gradually increase. It is difficult to findinformation requested. Therefore, the information on theWeb is expressed that is understandable, interpretable and

usable by computers to provide for finding the morequickly and easily.


6/35

Problem ?

To make semantic definition of the participatorsof NSDI

To implement the semantic matching betweenINSPIRE and GCM schemas


7/35

Semantic Definition-Projects

FinnONTO Project (2003-2012)-Semantic Computing

Research Group (SeCo)

SWING Project (2006-2009)-The Information Society Technologies (IST)

Program for Research, Technology Development & Demonstration under the Sixth

Framework Programme of the European Commission. (SINTEF-LFUI-UOM-IONIC-BRGM-

JSI-NUIG)

ACE-GIS Project (2002-2004)-the Five Framework Programme of the

European Commission (UOM, INESC-ID, E-Blana Enterprise Group, IONIC, UJI,SINTEF)


8/35

Semantic Definition- Academic Studies

Schade (2009). Ontology-Driven Translation of

Geospatial Data

Lemmens (2006). Semantic Interoperability of

Distributed Geo-Services

Dolbear vd, (2005). Semantic Interoperability

Between Topographic Data And A Flood Defence

Ontology


9/35

Semantic Web Requirements

The semantic definition of data and services.

The semantic annotation of data and services.

The semantic matching.


10/35

Methodology for Semantic Definition

1. The creation of syntactic definitions2. The selection transformation tools

3. The transformation the semantic web languages of

organization schema4. The selecting the suitable ontologies (Upper and domain)

5. The determining references between application and upper

level ontologies


11/35

Use CaseINSPIRE TN Road Schema-GML Application Schema (*.xsd)General Command of Mapping (GCM- Road Schema-UMLDiagram (*.vsd)For the semantic matching, the semantic definition is required.

?


12/35

The General Command of Mapping-GCM

(The National Mapping Agency of Turkey)

Why GCM?- Syntactic definition studies- One of the most main organizaton in NSDI

Standards of geographic domain in Turkey

- GCM FACC for 1/25000 scale map

- Large Scale Map and Map InformationProduction Regulation-FACC for 1/5000 scalemap


13/35

The Creation of Syntactic Definition

Syntactic Definition

- Feature and attribute definitions

- Database Schema, XML Schema


14/35

GCM Road Schema


15/35

- RDF (Resource Description Framework)- RDFS (RDF Schema)

- DAML+OIL (DARPA Agent Markup Language)+(Ontology

Interface Layer

- OWL (Web Ontology Language)

- WSML (Web Service Modeling Language)

Semantic Web Languages

Glten KARA, etin CMERT, 2011. Ulusal Konumsal Veri Altyaps iin Semantik Veri Tanmlama, Corafi Bilgi Sistemleri Kongresi,31.10.2011-04.11.2011, AKM, Antalya


16/35

The selection Semantik Web Language-Current tools

WSMT WSML, Protg OWL

-Expressivity providing of semantic web language

Concepts, relationships between concepts and constraints on them

the ability to express of each semantic web language is different.

But, semantic web languages use different logic languages.

Glten KARA, etin CMERT, 2011. Ulusal Konumsal Veri Altyaps iin Semantik Veri Tanmlama, Corafi Bilgi Sistemleri Kongresi,31.10.2011-04.11.2011, AKM, Antalya

Human

MaleFemale

Aye

hasAge 25

OWL uses disjoint classes in addition to RDFS

Selection the Semantic Web Language


17/35

The Selection Transformation Tools

TopBraidComposer, XSD2OWL, UML2OWL, Protege-UMLBackend, ArgoUML, XMLSpy, Umodel

DBtoOWL ?


18/35

The Selection Transformation Tools

Microsoft Office Visio 2010

Visual Paradigm for UML 9.0

XML (*.vdx)

UML2OWL

XMI (*.uml)

GCM Road Ontology

GCM Road Schema

OWL (*.owl)

Visio (*.vsd)

INSPIRE TN Schema (GML Application

Schema)XML Schema-OWL, DirectTranslationTopBraidComposer-commercial

GCM Road Schema (Microsoft Visio)

XML-UML-OWL , Indirect TranslationMicrosoft Visio Proffesional 2010-commercialVisual Paradigm for UML 9.0-commercialUML2OWL-open source


19/35

Transformation to the Semantic Web Language

Microsoft Visio

Protege

P d O t l Cl ifi ti


20/35

PhysicalObject

GeographicObject

Road

Highway

SpatialOperation

Intersection

Geometry

Polygon

GML OntologyDomain Ontology

DOLCE

Data Ontology

GCM:Road

Operation (Task) Ontology

LandslideArea

Landslide Ontology Operation Ontology

LandslideService

UPPER ONTOLOGY

LEVEL

DOMAIN ONTOLOGY

LEVEL

APPLICATION

ONTOLOGY LEVEL

DATA ONTOLOGY

LEVEL

Endurant

Particular

LineFeatureOperation

GCM:Topo25LineFeature

GCM:Highway_25m

GCM:Highway_25_50m

Proposed Ontology Classification


21/35

Selecting suitable ontology

How do I find the suitable ontology?

Ontology search engines and ontology libraries

Swoogle, Watson, Sindice,SWSE, Protege Ontology Library,DAML Ontology Library,.etc.

Currently, a standard tool do not exist for ontology selection

Upper level ontologies : DOLCE, SUMO, Cyc

We select DOLCE ontology because its dimension smaller others.


22/35

Semantic Definition

Semantic definition of an organization data is that

organization data is coded as understood by computers.

Organization

Schema

Feature

Definitions


23/35

Semantic Annotation

Semantic annotation is formal statementestablishing a link between concepts in ontologyand features in a data source.

Semantic Annotation

Application Ontology

Upper Ontology


24/35

Semantic Matching

If schemas thought as graph structure, SemanticMatchingcan be perceived as concepts of two graphnodes comparing semantically for determining thesimilarities between them.

Graph nodes may be concepts and attributes ofconcepts.

Semantic matching compares schema elements(concepts or attributes) semantically according to acommon schema, like ontologies.


25/35

Syntactic-Semantic Matching

Item Length Width

Highway 1000 km 20 m

Pathway . .

Type Length Breadth


Pathway . .

Syntactic

Matching

Semantic

Mapping

Road Roadbody

Item Length Width


Pathway . .

Type Length Breadth


Pathway . .

The existing datasets have different definitions of features.


26/35

S-MATCH

. In our schema matching scenario, we used S-MATCH software.

. It is generic semantic schema matching tool

. It takes two schemas (XML, OWL..) and returnssemantic relations between the nodes of theschemas using WordNet lexical database as an

external resource (Background Knowledgebase)


27/35

Why S-MATCH...

open-source

match schemas not only element level schema matchingbut also structure level schema matching

match schemas using both semantic and syntactictechniques

make use of a background knowledge base whenschema matching

Has extendable or changeable backgroundknowledgebase as our domain needs .

userfriendly


28/35

Implementation..

We implemented the semantic matchingbetween GCM-Road Ontology and INSPIRE-TNOntology with S-Match.

S-Match uses for semantic matching WordNet- Concepts

- Relationship between concepts

- Glosses of concepts


29/35

Implementation Architecture

S-MATCHMatch Results

WORDNET

INSPIRE TN Ontology GCM Road Ontology


30/35

Transformation Problems-Tools

Classes and attributes include Turkish characters

and / During the XML to UML (manually)

Software dependant data types(esriFieldTypeInteger) are not transformed. During

the UML to OWL UML2OWL is not correctly transform subtype

relationship.


31/35

Matching Results-LanguageGCM Road Ontology

1. GCM Road Ontology(Tur)-INSPIRE TN Ontology: Classes are matchedonly Thing class (Wordnet does not Turkish language)

Subtype relationship Without Subtype relationship


32/35

Subtype relationship Without Subtype relationship

Matching Results-LanguageGCM Road Ontology(Eng)-INSPIRE TN Ontology: GCM Road ontology istranslated English language.

First hierarchy includes Subtype relationship. GCM Road > INSPIRE RoadSecond hierarchy do not includes Subtype relationship. INSPIRE Road > GCM Road


33/35

Future Works and Conclusion

We proposed a methodology for semanticdefinition of an organization data in NSDI

We are planning to extend WordNet thatcontains Turkish spatial concepts, attributes andrelations.

We are performed semantic matching in theclass level. We plan to make works aboutmapping of schema attributes.


34/35

Future Works and Conclusion

Tools should be developedfor the transformation of software dependant datatypes (esriFieldTypeInteger),

for the correction of names of classes andattributes

for transformation from syntactic definition tosemantic definition


35/35

Thank you for your attention.

INSPIRE 2012-Gülten Kara

Documents

Transcript of INSPIRE 2012-Gülten Kara