A SEMANTIC GEO-CATALOGUE IMPLEMENTATION FOR AIMPLEMENTATION FOR A REGIONAL SDIREGIONAL SDI

P Shvaiko Informatica TrentinaP. Shvaiko, Informatica TrentinaA. Ivanyukovich, Trient Consulting Group

L. Vaccari, EC DG JRCV M lt U i it f T tV. Maltese, University of Trento

F. Farazi, University of Trento

INSPIRE conference - June 24th 2010 Feroz Farazi – University of Trento

O lOutline

IntroductionIntroduction Geo-catalogue implementationSemantic extension for the geo catalogueSemantic extension for the geo-catalogueConclusions

INSPIRE conference - June 24th 2010 Feroz Farazi – University of Trento

S T

Introduction Implementation Semantic extension Conclusions

SDI in Trentino

INSPIRE conference - June 24th 2010 Feroz Farazi – University of Trento

S f G

Introduction Implementation Semantic extension Conclusions

Services for GI access

Services for citizensServices for citizensServices for companiesServices for research centersServices for research centersServices for public administrations

INSPIRE conference - June 24th 2010 Feroz Farazi – University of Trento

G l l h

Introduction Implementation Semantic extension Conclusions

Geo-portal > geo-catalogue home page

INSPIRE conference - June 24th 2010 Feroz Farazi – University of Trento

A h

Introduction Implementation Semantic extension Conclusions

Architecture - I

The Geo catalogue consists of the following main modules:The Geo-catalogue consists of the following main modules:

Metadata managementMetadata managementAdd/Import/Modify/Delete/Permissions/State management (workflow)

System configurationSystem configurationHarvesting/Localization/Thesaurus management

User & Group managementUser & Group managementImport users/Role mgmt/Group mgmt/Ownership transfer

Semantic matching engineSemantic matching engineIntelligent query expansion

INSPIRE conference - June 24th 2010 Feroz Farazi – University of Trento

A h

Introduction Implementation Semantic extension Conclusions

Architecture - II

INSPIRE conference - June 24th 2010 Feroz Farazi – University of Trento

l

Introduction Implementation Semantic extension Conclusions

Implementation notes

Workflow model for metadata management:Workflow model for metadata management:Notion of states – Draft, Review and Published…

Metadata versioningMetadata validation using configurable rules:Metadata validation using configurable rules:

Min field length (in words)Stop words…

Metadata quality controlSemaphore for fast quality check

RNDT v1.0 - compliant implementation for Vector, Raster and SService metadata templates

INSPIRE conference - June 24th 2010 Feroz Farazi – University of Trento

S

Introduction Implementation Semantic extension Conclusions

Semantic extension: overview

Problems with current tools for search (e g keyword search)Problems with current tools for search (e.g. keyword search)Different terminology used ➭ low recallPoor expressivity ➭ low precisionPoor expressivity low precision

Intelligent query expansion to improve user experienceg q y p p pbased on Semantic Matching techniques: the S-match system

INSPIRE conference - June 24th 2010 Feroz Farazi – University of Trento

S h d S h

Introduction Implementation Semantic extension Conclusions

Semantic matching and S-MatchSemantic matching is a technique for identifying semanticSemantic matching is a technique for identifying semantic correspondences between ontologiesA set of matches is called alignmentg

A C RIVER⊒ ⊑ MORE SPECIFICA C RIVERSTREAM ⊒

⊒⊑⊒≡

MORE GENERAL

EQUIVALENT TOB DTRENTO POVO

⊒ ⊥Q

DISJOINT FROM

S-Match is a tool for semantic matching developed at the University of TrentoyS-Match is an open source tool, available at: http://semanticmatching.org/p g g

INSPIRE conference - June 24th 2010 Feroz Farazi – University of Trento

f S h h l

Introduction Implementation Semantic extension Conclusions

Integration of S-Match into the geo-catalogue

S Match as web service advantages:S-Match as web-service - advantages:Separate VM usage for:

Performance optimizationPerformance optimizationFlexible memory management

Simplified updates in production environment:Simplified updates in production environment:Ontology updates

INSPIRE conference - June 24th 2010 Feroz Farazi – University of Trento

Th l f T h d l

Introduction Implementation Semantic extension Conclusions

The ontology of Trentino: methodologyObjective: Create an ontology that reflects the specificity of j gy p yTrentino and repects the canons of the analytico-syntheticapproach for the generation of a faceted lightweight ontology

Analysis and disambiguation of the classesAnalysis and disambiguation of the classesHierarchical reorganization: semantic relations

Available data: The ontological analysis was accomplishedAvailable data: The ontological analysis was accomplished on the top 45 categories of place names (numbered from P110 to P650) provided by PAT

The generated ontology includes 5 facets:AntiquityAntiquityBody of water (rivers, lakes, …)Geological formation (mountains, valleys, …)Facility (roads, railways, shelters, …)Administrative Division (provinces, municipalities, wards, …)

INSPIRE conference - June 24th 2010 Feroz Farazi – University of Trento

Th l f T d f

Introduction Implementation Semantic extension Conclusions

The ontology of Trentino: Body of water

Categories from P210 to P251Categories from P210 to P251

P210 Corsi dacqua/laghi (1 ord ) Body of water (Idrografia)P210 Corsi dacqua/laghi (1 ord.)P220 Corsi dacqua/laghi (2 ord.)P230 Corsi

Body of water (Idrografia)

Lake (Lago)d’acqua/Canali/Fosse/Cond. forz./Laghi (3 ord.)P240 Corsi

( g )Group of lakes (Gruppo di laghi)Stream (Corso d’acqua)

Ri er (Fi me)P240 Corsi d’acqua/Canali/Fosse/Cond. forz./Laghi (>3 ord.-25.000)

River (Fiume)Rivulet (Torrente)

Spring (Sorgente)P241 Corsi d’acqua/Canali/Fosse/Cond. forz./Laghi (>3 ord.)

p g ( g )Waterfall (Cascata)

Cascade (Cascatina)Canal (Canale)g ( )

P250 Sorg-Casc-Font-Presa importantiP251 S C F t P i i

Canal (Canale)

P251 Sorg-Casc-Font-Presa minori INSPIRE conference - June 24th 2010 Feroz Farazi – University of Trento

l f S h

Introduction Implementation Semantic extension Conclusions

Usage examples of S-MatchQuery Query expansion with S-Match

watercourse Rivulet, Stream, River

falls Cascade, Waterfall

water Rivulet, Waterfall, Cascade, River, Body of water, Stream, Spring, Canal, Group of lakes, Lake

l i N l l i M i Hi hl d Gl i M ielevation Natural elevation, Mountain, Highland, Glacier, Mountain range, Peak, Hill

installation Milestone, Hut, Farm, Highway, Railway, Road, Street, , , , g y, y, , ,Transportation system, Provincial Road, Facility, Shelter

ice mass Glacier

district Administrative division, Province, Municipality, Ward, Populated place

transportation facility Transportation system Road Street Provincial Road Milestonetransportation facility Transportation system, Road, Street, Provincial Road, Milestone, Railway, Highway

cliff

INSPIRE conference - June 24th 2010 Feroz Farazi – University of Trento

C l

Introduction Implementation Semantic extension Conclusions

Conclusions We found it useful to realize our geo-catalogue service from g gthe available open-source implementation (GeoNetwork) conforming to the INSPIRE Directive and extend it with semantic capabilitiespIt was adapted based on the available technological infrastructure (the geo-portal of Trentino) and specificities of the implementation rules further constrained at the nationalthe implementation rules further constrained at the national level, e.g., by RNDT.

Future work includes an extensive empirical evaluation of the semantic extension of the geo-catalogue. The expected impact here includes:The expected impact here includes:

Better automation ➭ time and cost savingsI can find more rapidly the data that I am interested inI can find more rapidly the data that I am interested inBetter accuracy ➭ more precise decisionsExhaustive results for search of data and services based on the available resourceavailable resource

INSPIRE conference - June 24th 2010 Feroz Farazi – University of Trento

THANK YOUTHANK YOU FOR YOUR ATTENTION AND INTEREST

INSPIRE conference - June 24th 2010 Feroz Farazi – University of Trento