Abstract image of a woman sitting on books and studying on a laptop

Instance-Based Ontological Knowledge Acquisition

Lihua Zhao, profile picture
Lihua Zhao

This document presents a framework for instance-based ontological knowledge acquisition, focusing on semi-automatic ontology integration of heterogeneous datasets in linked open data. It discusses techniques such as machine learning for ontology schema extraction and graph-based ontology integration to address ontology heterogeneity and class/property grouping. The framework's effectiveness is evaluated through experiments involving various datasets, highlighting its potential for creating a global ontology.

TechnologyEducation
Related topics:
Semantic Web • Linked Data
1 of 33
Downloaded 17 times
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Instance-Based Ontological Knowledge Acquisition The Graduate University for Advanced Studies (SOKENDAI) National Institute of Informatics Lihua Zhao & Ryutaro Ichise ESWC2013, Montpellier, France, 28th May, 2013
Outline Introduction Related Work Semi-automatic Ontology Integration Framework Graph-Based Ontology Integration Machine-Learning-Based Ontology Schema Extraction Ontology Merger Experiments Conclusion and Future Work Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 2
Introduction Linked Open Data (LOD) Machine-readable and interlinked at instance-level. 295 data sets, 31 billion RDF triples (as of Sep. 2011). Around 504 million owl:sameAs links. 7 domains (cross-domain, geographic, media, life sciences, government, user-generated content, and publications). World Fact- book John Peel (DBTune) Pokedex Pfam US SEC (rdfabout) Linked LCCN Europeana EEA IEEE ChEMBL Semantic XBRL SW Dog Food CORDIS (FUB) AGROVOC Openly Local Discogs (Data Incubator) DBpedia yovisto Tele- graphis tags2con delicious NSF Medi Care Brazilian Poli- ticians dotAC ERA Open Cyc Italian public schools UB Mann- heim JISC Moseley Folk Semantic Tweet OS GTAA totl.net OAI Portu- guese DBpedia LOCAH KEGG Glycan CORDIS (RKB Explorer) UMBEL Affy- metrix riese business. data.gov. uk Open Data Thesau- rus Geo Linked Data UK Post- codes Smart Link ECCO- TCP UniProt (Bio2RDF) SSW Thesau- rus RDF ohloh Freebase London Gazette Open Corpo- rates Airports GEMET P20 TCM Gene DIT Source Code Ecosystem Linked Data OMIM Hellenic FBD Data Gov.ie Music Brainz (DBTune) data.gov.uk intervals LODE Climbing SIDER Project Guten- berg Music Brainz (zitgist) ProDom HGNC SMC Journals Reactome National Radio- activity JP legislation data.gov.uk AEMET Product Types Ontology Linked User Feedback Revyu Gene Ontology NHS (En- AKTing) URI Burner DB Tropes Eurécom ISTAT Immi- gration Lichfield Spen- ding Surge Radio Euro- stat (FUB) Piedmont Accomo- dations New York Times Klapp- stuhl- club EUNIS Bricklink reegle CO2 Emission (En- AKTing) Audio Scrobbler (DBTune) GovTrack GovWILD ECS South- ampton EPrints KEGG Reaction Linked EDGAR (Ontology Central) LIBRIS Open Library KEGG Drug research. data.gov. uk VIVO Cornell UniRef WordNet (RKB Explorer) Cornetto medu- cator DDC Deutsche Bio- graphie Wiki Ulm NASA (Data Incu- bator) BBC Music Drug Bank Turismo de Zaragoza Plymouth Reading Lists education. data.gov. uk KISTI Uni Pathway Eurostat (Ontology Central) OGOLOD Twarql Music Brainz (Data Incubator) Geo Names Pub Chem Italian Museums Good- win Family flickr wrappr Eurostat Thesau- rus W Open Library (Talis) LOIUS Linked GeoData Linked Open Colors WordNet (VUA) patents. data.gov. uk Greek DBpedia Sussex Reading Lists Metoffice Weather Forecasts GND LinkedCT SISVU transport. data.gov. uk Didac- talia dbpedia lite BNB Ontos News Portal LAAS Product DB iServe Recht- spraak. nl KEGG Com- pound Geo Species VIVO UF Linked Sensor Data (Kno.e.sis) lobid Organi- sations LEM Linked Crunch- base FTS Ocean Drilling Codices Janus AMP ntnusc Weather Stations Amster- dam Museum lingvoj Crime (En- AKTing) Course- ware PubMed ACM BBC Wildlife Finder Calames Chronic- ling America data- open- ac- uk Open Election Data Project Slide- share2RDF Finnish Munici- palities OpenEI MARC Codes List VIVO Indiana Hellenic PD LCSH FanHubz bible ontology IdRef Sudoc KEGG Enzyme NTU Resource Lists PRO- SITE Linked Open Numbers Energy (En- AKTing) Roma Open Calais data bnf.fr lobid Resources IRIT theses. fr LOV Rådata nå! Daily Med Taxo- nomy New- castle Google Art wrapper Poké- pédia EURES BibBase RESEX STITCH PDB EARTh IBM Last.FM artists (DBTune) YAGO ECS (RKB Explorer) Event Media STW my Experi- ment BBC Program- mes NDL subjects Taxon Concept Pisa KEGG Pathway UniParc Jamendo (DBtune) Popula- tion (En- AKTing) Geo- WordNet RAMEAU SH UniSTS Mortality (En- AKTing) Alpine Ski Austria DBLP (RKB Explorer) Chem2 Bio2RDF MGI DBLP (L3S) Yahoo! Geo Planet GeneID RDF Book Mashup El Viajero Tourism Uberblic Swedish Open Cultural Heritage GESIS data dcs Last.FM (rdfize) Ren. Energy Genera- tors Sears RAE2001 NSZL Catalog Homolo- Gene Ord- nance Survey TWC LOGD Disea- some EUTC Produc- tions PSH WordNet (W3C) semantic web.org Scotland Geo- graphy Magna- tune Norwe- gian MeSH SGD Traffic Scotland statistics. data.gov. uk Crime Reports UK UniProt US Census (rdfabout) Man- chester Reading Lists EU Insti- tutions PBAC VIAF UN/ LOCODE Lexvo Linked MDB ESD stan- dards reference. data.gov. uk t4gm info Sudoc ECS South- ampton ePrints Classical (DB Tune) DBLP (FU Berlin) Scholaro- meter St. Andrews Resource Lists NVD Fishes of TexasScotland Pupils & Exams RISKS gnoss DEPLOY InterPro Lotico Ox Points Enipedia ndlna Budapest CiteSeer Media Geographic Publications User-generated content Government Cross-domain Life sciences As of September 2011 Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 3
Motivation Figure: Interlinked Instances of “France”. Problems when access to several data sets: Ontology Heterogeneity Problem Map related ontology classes and properties. Ontology similarity matching on the SameAs graph patterns. Difficulty in Identifying Core Ontology Schemas Retrieve frequently used core ontology classes and properties. Machine learning for core ontology schema extraction. Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 4
Related Work Find useful attributes from frequent graph patterns using a supervised machine learning method. [Le, 2010] Only for geographic data and no discussion about the features. A debugging method for mapping lightweight ontologies with machine learning method. [Meilicke, 2008] Limited to the expressive lightweight ontologies. Construct intermediate-layer ontology by analyzing concept coverings. [Parundekar, 2012] Only for specific domains and limited between two resources. Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 5
Semi-automatic Ontology Integration Framework Construct a global ontology by integrating heterogeneous ontologies of the Linked Open Data. Graph-Based Ontology Integration [Zhao, et al., 2012] Group related classes and properties. Machine-Learning-Based Ontology Schema Extraction Extract frequent core classes and properties. Ontology Merger Merge extracted ontology classes and properties. Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 6
Semi-automatic Ontology Integration Framework Construct a global ontology by integrating heterogeneous ontologies of the Linked Open Data. Graph-Based Ontology Integration [Zhao, et al., 2012] Group related classes and properties. Machine-Learning-Based Ontology Schema Extraction Extract frequent core classes and properties. Ontology Merger Merge extracted ontology classes and properties. Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 7
Graph-based Ontology Integration Extract graph patterns from interlinked instances to discover related ontology classes and predicates. Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 8
STEP 1: Graph Pattern Extraction SameAs Graph: An undirected SameAs Graph SG = (V , E, I), where V : a set of vertices (the labels of data sets). E ⊆ V × V : a set of sameAs edges. I: a set of URIs of the interlinked SameAs Instances. Example: SGFrance = (VFrance, EFrance, IFrance). VFrance = {M, D, G, N} EFrance = {(D, G), (D, N), (G, M), (G, N)} IFrance = {mdb-country:FR, db:France, geo:3017382, nyt:67...21}. Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 9
STEP 2: <Predicate, Object> Collection <Predicate, Object> (PO) pairs and types for SGFrance Predicate Object Type rdf:type db-onto:Country Class rdfs:label “France”@en String foaf:name “France”@en String foaf:name “R´epublique fran¸caise”@en String db-onto:wikiPageExternalLink http://us.franceguide.com/ URI db-prop:populationEstimate 65447374 Number . . . . . . . . . . . . . . . . . . geo-onto:name France String geo-onto:alternateName “France”@en String geo-onto:featureCode geo-onto:A.PCLI Class geo-onto:population 64768389 Number . . . . . . . . . . . . . . . . . . rdf:type mdb:country Class mdb:country name France String mdb:country population 64094000 Number rdfs:label France (Country) String . . . . . . . . . . . . . . . . . . rdf:type skos:Concept Class skos:inScheme nyt:nytd geo Class skos:prefLabel “France”@en String nyt-prop:first use 2004-09-01 Date Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 10
STEP 3: Related Classes and Properties Grouping Related Classes Grouping (Leaf nodes) Tracking subsumption relations from SameAs graphs. < C1 owl:subClassOf C2 > < C1 skos:inScheme C2 > Example: SGFrance Related Classes → {db-onto:Country, geo-onto:A.PCLI, mdb:country, nyt:nytd geo } Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 11
STEP 3: Related Classes and Properties Grouping Related Properties Grouping Exact matching for creating initial sets of PO pairs S1, S2, . . . , Sk . Similarity matching on the initial sets of PO pairs. Sim(POi , POj ) = ObjSim(POi , POj ) + PreSim(POi , POj ) 2 ObjSim(POi , POj ) =    1 − |OPOi −OPOj | OPOi +OPOj if OPO is Number StrSim(OPOi , OPOj ) if OPO is String PreSim(POi , POj ) = WNSim(TPOi , TPOj ) StrSim(OPOi , OPOj ): Average of 3 string-based similarity measures. WNSim(TPOi , TPOj ): Average of 9 WordNet-based similarity measures. Refine sets of PO pairs according to rdfs:domain. Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 12
STEP 4: Aggregation of All Graph Patterns Aggregate the integrated classes and properties from all the extracted graph patterns. Select A Term for Each Set ex-onto:ClassTerm ex-onto:propTerm Construct Relations ex-prop:hasMemberClasses <class, ex-prop:hasMemberClasses, ex-onto:ClassTerm> ex-prop:hasMemberDataTypes <property, ex-prop:hasMemberDataTypes, ex-onto:propTerm> Construct A Preliminary Integrated Ontology Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 13
STEP 5: Manual Revision Manually revise the preliminary integrated ontology. Terms of the integrated classes and properties: Choose a proper term for each group of classes or properties. Groups of related classes or properties: Correct wrong grouping. Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 14
Semi-automatic Ontology Integration Framework Construct a global ontology by integrating heterogeneous ontologies of the Linked Open Data. Graph-Based Ontology Integration [Zhao, et al., 2012] Group related classes and properties. Machine-Learning-Based Ontology Schema Extraction Extract frequent core classes and properties. Ontology Merger Merge extracted ontology classes and properties. Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 15
Machine-Learning-Based Ontology Schema Extraction Top-level classes and core properties are necessary. Decision Table Retrieves core properties in each data set. Belongs to rule-based machine learning with simple hypothesis. Retrieves a subset of properties that are important for describing instances in a data set. Apriori Retrieves core properties in the instances of a specific top-level class. Belongs to association rule mining. Finds a set of properties, whose support is greater than the user-defined minimum support. Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 16
Decision Table Retrieve top-level classes and core properties that are important for describing instances in a data set. Collect top-level classes. Filter out infrequent properties. Convert each instance for the Decision Table algorithm. weight(prop1, inst), weight(prop2, inst), ... weight(propn, inst), class PF-IIF (Property Frequency - Inverse Instance Frequency) weight(prop, inst) = pf (prop, inst) × iif (prop, D) pf (prop, inst) = the frequency of prop in inst. iif (prop, D) = log |D| |instprop| instprop: an instance that contains the property prop. Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 17
Apriori Retrieve top-level classes and frequent core properties that are important for describing instances in a specific class. Collect top-level classes. Filter out infrequent properties. Convert each instance of top-level class c for the Apriori algorithm. [prop1, prop2, ..., propn] Define minimum support and confidence metric. Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 18
Semi-automatic Ontology Integration Framework Construct a global ontology by integrating heterogeneous ontologies of the Linked Open Data. Graph-Based Ontology Integration [Zhao, et al., 2012] Group related classes and properties. Machine-Learning-Based Ontology Schema Extraction Extract frequent core classes and properties. Ontology Merger Merge extracted ontology classes and properties. Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 19
Ontology Merger Graph-Based Ontology Integration outputs a Preliminary Integrated Ontology. For the ontology classes and properties retrieved from Machine-Learning-Based Approach: If Class c ∈ Preliminary Integrated Ontology, add < ex-onto:ClassTermnew , ex-prop:hasMemberClasses, c >. For each Property prop retrieved from top-level class c using Apriori, add a triple < prop, rdfs:domain, c >. Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 20
Experiments Data Sets Graph-Based Ontology Integration Decision Table Apriori Comparison of Integrated Ontology Case Studies Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 21
Data Sets DBpedia (v3.6): cross-domain, 3.5 million things, 8.9 million URIs. Geonames (v2.2.1): geographical domain, 7 million URIs. NYTimes: media domain, 10,467 subject news. LinkedMDB: media domain, 0.5 million entities. Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 22
Data Sets - Machine Learning Data Set Instances Selected Class Top-level Property Selected Instances Class Property DBpedia 3,708,696 64,460 241 28 1385 840 Geonames 7,480,462 45,000 428 9 31 21 NYTimes 10,441 10,441 5 4 8 7 LinkedMDB 694,400 50,000 53 10 107 60 Selected Instances Randomly select instances per class: DBpedia (5000), Geonames(3000), NYTimes(All), LinkedMDB(3000) Top-level Classes Ontology-based data set: Use subsumption relations. Without ontology: Use categories. Selected Properties With frequency threshold θ as √ n, where n is the total number of instances in the data set. Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 23
Graph-Based Ontology Integration 13 graph patterns Frequent graph patterns: GP1, GP2, GP3 N,G,D: GP4, GP5, GP7, GP8 N,M,D: GP6 M,G,D: GP9 M,D,N,G: GP10, GP11, GP12, GP13 13 graph patterns. 97 classes into 48 groups. 357 properties into 38 groups. Retrieved related classes and properties by analyzing graph patterns. [Zhao, I-Semantics2012] Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 24
Evaluation of Machine Learning Approaches Evaluate the Decision Table and Apriori algorithm. Evaluation of Decision Table Evaluate whether the retrieved sets of properties are important for describing instances by testing if they can be used to distinguish different types of instances in the data set. Evaluation of Apriori Analyze the performance of Apriori algorithm in each data set with examples of retrieved sets of properties. Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 25
Decision Table Data Set Precision Recall F-Measure Selected Properties DBpedia 0.892 0.821 0.837 53 Geonames 0.472 0.4 0.324 10 NYTimes 0.795 0.792 0.785 5 LinkedMDB 1 1 1 11 Core properties are evaluated by predicting classes of instances (10-fold). 11 properties from LinkedMDB can correctly identify class of an instance. DBpedia and NYTimes performs good with selected properties. 10 properties from Geonames are commonly used for all types of classes. Examples of retrieved core properties. DBpedia: db-prop:city, db-prop:debut, db-onto:formationYear,etc. Geonames: geo-onto:alternateName, geo-onto:countryCode, etc. NYTimes: nyt:latest use, nyt:topicPage, wgs84 pos:long, etc. LinkedMDB: mdb:director directorid, mdb:writer writerid, etc. Retrieved top-level classes and core properties in each data set. Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 26
Apriori Examples of retrieved core properties with Apriori Algorithm. Data Set Class Properties DBpedia db:Event db-onto:place, db-prop:date, db-onto:related/geo. db:Species db-onto:kingdom, db-onto:class, db-onto:family. db:Person foaf:givenName, foaf:surname, db-onto:birthDate. Geonames geo:P geo-onto:alternateName, geo-onto:countryCode. geo:R wgs84 pos:alt, geo-onto:name, geo-onto:countryCode. NYTimes nyt:nytd geo wgs84 pos:long. nyt:nytd des skos:scopeNote. LinkedMDB mdb:actor mdb:performance, mdb:actor name, mdb:actor netflix id. mdb:film mdb:director, mdb:performane, mdb:actor, dc:date. DBpedia and LinkedMDB: Retrieved unique properties. Geonames and NYTimes: Retrieved commonly used properties only. Automatically added missing domain information: < prop, rdfs : domain, classtop >. Retrieved frequent core properties in each top-level class. Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 27
Comparison of Integrated Ontology Previous Work Machine Learning Current Work Graph-Based Decision Apriori Integrated Integration Table Ontology Class 97 50 (38 new) 50 (38 new) 135 (38 new) Property 357 79 (49 new) 119(80 new) 453 (96 new) Previous Work: 97 classes in 49 groups, 357 properties in 38 groups. Current Work: 135 classes in 87 groups, 453 properties in 97 groups. Apriori retrieves more properties than Decision Table. 33 new properties are found with both Apriori and Decision Table. Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 28
Case Studies I Find Missing Links of Islands with Integrated Ontology SELECT DISTINCT ?geo ?db ?string where { ?geo geo-onto:featureCode geo-onto:T.ISL. ?geo ?gname ?string. ex-onto:name ex-prop:hasMemberDataTypes ?gname. ?db rdf:type db-onto:Island. ex-onto:name ex-prop:hasMemberDataTypes ?dname. ?db ?dname ?string. } Retrieved 509 links, including 218 existing SameAs links: 97 existing links from DBpedia to Geonames. 211 links from Geonames to DBpedia. 90 bidirectional links between DBpedia and Geonames. Discovered 291 missing links with the integrated ontology using exact matching on the labels of instances. Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 29
Case Studies II Predicates grouped in ex-prop:birthDate Property Number of Instances rdfs:domain db-onto:birthDate 287,327 db-onto:Person db-prop:datebirth 1,675 N/A db-prop:dateofbirth 87,364 N/A db-prop:dateOfBirth 163,876 N/A db-prop:born 34,832 N/A db-prop:birthdate 70,630 N/A db-prop:birthDate 101,121 N/A Suggest “db-onto:birthDate” as the standard property because it has rdfs:domain definition has the highest usage in the DBpedia instances. Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 30
Case Studies III Give me all the cities with more than 10,000,000 inhabitants. Standard Query Query with the Integrated Ontology SELECT DISTINCT ?uri ?string SELECT DISTINCT ?uri ?string WHERE { WHERE { ?uri rdf:type db-onto:City. ?uri rdf:type db-onto:City. ex-onto:population ex-prop:hasMemberDataTypes ?prop. ?uri db-prop:populationTotal ?inhabitants. ?uri ?prop ?inhabitants. FILTER (?inhabitants > 10000000). FILTER (?inhabitants > 10000000). OPTIONAL { ?uri rdfs:label ?string. OPTIONAL { ?uri rdfs:label ?string. FILTER (lang(?string) = ’en’) }} FILTER (lang(?string) = ’en’) }} A SPARQL example from QALD-1 Open Challenge. Standard query: 9 cities. Query with the integrated ontology: 20 cities. Help QA systems for finding more related answers with simple queries. Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 31
Conclusion and Future Work Conclusion Semi-automatic ontology integration framework Graph-Based Ontology Integration. Ontology similarity matching on SameAs graph patterns. Retrieve related ontology classes and properties. Machine-Learning-Based Ontology Schema Extraction Decision Table and Apriori. Extract top-level classes and core properties. Ontology Merger Find missing links, detect misuses of ontologies, and access various data sets with integrated ontology. Future Work Automatically detect and revise mistakes in ontology merger. Automatically detect ranges and domains of properties. Test our framework with more LOD data sets. Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 32
Thank you! Questions? Lihua Zhao, lihua@nii.ac.jp Ryutaro Ichise, ichise@nii.ac.jp Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 33

More Related Content

PDF
Mid-Ontology Learning from Linked Data @JIST2011
Lihua Zhao
 
PDF
Interlinking educational data to Web of Data (Thesis presentation)
Enayat Rajabi
 
PPTX
The Landscape of Ontology Reuse in Linked Data - OEDW2012
María Poveda Villalón
 
PPTX
Pattern-based Acquisition of Scientific Entities from Scholarly Article Title...
Jennifer D'Souza
 
PDF
What makes a linked data pattern interesting?
Szymon Klarman
 
PDF
Perspectives on mining knowledge graphs from text
Jennifer D'Souza
 
PDF
VDOS2013-Zhe-Slides
Zhe (Henry) He
 
PPT
2011linked science4mccuskermcguinnessfinal
Deborah McGuinness
 
Mid-Ontology Learning from Linked Data @JIST2011
Lihua Zhao
 
Interlinking educational data to Web of Data (Thesis presentation)
Enayat Rajabi
 
The Landscape of Ontology Reuse in Linked Data - OEDW2012
María Poveda Villalón
 
Pattern-based Acquisition of Scientific Entities from Scholarly Article Title...
Jennifer D'Souza
 
What makes a linked data pattern interesting?
Szymon Klarman
 
Perspectives on mining knowledge graphs from text
Jennifer D'Souza
 
VDOS2013-Zhe-Slides
Zhe (Henry) He
 
2011linked science4mccuskermcguinnessfinal
Deborah McGuinness
 

What's hot (19)

PDF
Using Substitutive Itemset Mining Framework for Finding Synonymous Properties...
Agnieszka Ławrynowicz
 
DOCX
Advances of neural networks in 2020
kevig
 
PDF
Connecting life sciences data at the European Bioinformatics Institute
Connected Data World
 
PPTX
Ontologies neo4j-graph-workshop-berlin
Simon Jupp
 
PDF
OntoMath digital ecosystem
Alik Kirillovich
 
DOCX
Ontology based clustering algorithms
Ikutwa
 
PPTX
Connected Data for Machine Learning | Paul Groth
Connected Data World
 
PPTX
schema.org and biomedical ontologies
Simon Jupp
 
PPT
DB-IR-ranking
FELIX75
 
PPTX
Semantics as a service at EMBL-EBI
Simon Jupp
 
PDF
RuleML2015: FOWLA, a federated architecture for ontologies
RuleML
 
PPTX
Facilitating semantic alignment.-biohackathon-jupp
Simon Jupp
 
PPTX
20130622 okfn hackathon t2
Seonho Kim
 
PDF
Survey on Parallel/Distributed Search Engines
Yu Liu
 
PDF
AN IMPROVED TECHNIQUE FOR DOCUMENT CLUSTERING
International Journal of Technical Research & Application
 
PPT
DB and IR Integration
Marco A Torres
 
PPTX
Building a repository of biomedical ontologies with Neo4j
Simon Jupp
 
PPT
Importing life science at a into Neo4j
Simon Jupp
 
PDF
Linked Open (Geo)Data and the Distributed Ontology Language – a perfect match
Christoph Lange
 
Using Substitutive Itemset Mining Framework for Finding Synonymous Properties...
Agnieszka Ławrynowicz
 
Advances of neural networks in 2020
kevig
 
Connecting life sciences data at the European Bioinformatics Institute
Connected Data World
 
Ontologies neo4j-graph-workshop-berlin
Simon Jupp
 
OntoMath digital ecosystem
Alik Kirillovich
 
Ontology based clustering algorithms
Ikutwa
 
Connected Data for Machine Learning | Paul Groth
Connected Data World
 
schema.org and biomedical ontologies
Simon Jupp
 
DB-IR-ranking
FELIX75
 
Semantics as a service at EMBL-EBI
Simon Jupp
 
RuleML2015: FOWLA, a federated architecture for ontologies
RuleML
 
Facilitating semantic alignment.-biohackathon-jupp
Simon Jupp
 
20130622 okfn hackathon t2
Seonho Kim
 
Survey on Parallel/Distributed Search Engines
Yu Liu
 
AN IMPROVED TECHNIQUE FOR DOCUMENT CLUSTERING
International Journal of Technical Research & Application
 
DB and IR Integration
Marco A Torres
 
Building a repository of biomedical ontologies with Neo4j
Simon Jupp
 
Importing life science at a into Neo4j
Simon Jupp
 
Linked Open (Geo)Data and the Distributed Ontology Language – a perfect match
Christoph Lange
 
Ad

Similar to Instance-Based Ontological Knowledge Acquisition (20)

PPTX
Discovering Alignments in Ontologies of Linked Data
Craig Knoblock
 
PDF
Knowledge Engineering: Semantic web, web of data, linked data
Franck Michel
 
PDF
Ontology Engineering
Alessandro Adamou
 
PPTX
Leveraging Linked Data to Infer Semantic Relations within Structured Sources
Mohsen Taheriyan
 
PDF
Mapping Lo Dto Proton Revised [Compatibility Mode]
Mariana Damova, Ph.D
 
PPTX
Semantic Integration for Heterogeneous Domain-specific Information: The NIF Case
Neuroscience Information Framework
 
PDF
Reasoning with Big Knowledge Graphs: Choices, Pitfalls and Proven Recipes
Ontotext
 
PPTX
Knowledge Graph Introduction
Sören Auer
 
PDF
bridging formal semantics and social semantics on the web
Fabien Gandon
 
PDF
Harnessing Linked Knowledge Sources for Topic Classification in Social Media
Amparo Elizabeth Cano Basave
 
PPTX
Ontology integration - Heterogeneity, Techniques and more
Adriel Café
 
PPTX
sOCIAL NETWORK ANALYSIS AND ONTOLOGIES A VIEW
syedanwarf
 
PDF
Danbri Drupalcon Export
Drupalcon Paris
 
PDF
WEB ONTOLOGY LANGUAGE: OWL
Tochukwu Udeh
 
PDF
Owl assignment udeh
Tochukwu Udeh
 
PPTX
A Graph-based Approach to Learn Semantic Descriptions of Data Sources
Pedro Szekely
 
PPTX
A Graph-Based Approach to Learn Semantic Descriptions of Data Sources
Mohsen Taheriyan
 
PPTX
Presentation of Profiling Similarity Links in LOD @ DesWEB, ICDE 2016
Blerina Spahiu
 
PPTX
Wi2015 - Clustering of Linked Open Data - the LODeX tool
Laura Po
 
PPTX
Release webinar: Sansa and Ontario
BigData_Europe
 
Discovering Alignments in Ontologies of Linked Data
Craig Knoblock
 
Knowledge Engineering: Semantic web, web of data, linked data
Franck Michel
 
Ontology Engineering
Alessandro Adamou
 
Leveraging Linked Data to Infer Semantic Relations within Structured Sources
Mohsen Taheriyan
 
Mapping Lo Dto Proton Revised [Compatibility Mode]
Mariana Damova, Ph.D
 
Semantic Integration for Heterogeneous Domain-specific Information: The NIF Case
Neuroscience Information Framework
 
Reasoning with Big Knowledge Graphs: Choices, Pitfalls and Proven Recipes
Ontotext
 
Knowledge Graph Introduction
Sören Auer
 
bridging formal semantics and social semantics on the web
Fabien Gandon
 
Harnessing Linked Knowledge Sources for Topic Classification in Social Media
Amparo Elizabeth Cano Basave
 
Ontology integration - Heterogeneity, Techniques and more
Adriel Café
 
sOCIAL NETWORK ANALYSIS AND ONTOLOGIES A VIEW
syedanwarf
 
Danbri Drupalcon Export
Drupalcon Paris
 
WEB ONTOLOGY LANGUAGE: OWL
Tochukwu Udeh
 
Owl assignment udeh
Tochukwu Udeh
 
A Graph-based Approach to Learn Semantic Descriptions of Data Sources
Pedro Szekely
 
A Graph-Based Approach to Learn Semantic Descriptions of Data Sources
Mohsen Taheriyan
 
Presentation of Profiling Similarity Links in LOD @ DesWEB, ICDE 2016
Blerina Spahiu
 
Wi2015 - Clustering of Linked Open Data - the LODeX tool
Laura Po
 
Release webinar: Sansa and Ontario
BigData_Europe
 
Ad

Recently uploaded (20)

PDF
Lung cancer patients survival prediction using outlier detection and optimize...
IAESIJAI
 
PDF
Dell Pro Micro: Speed customer interactions, patient processing, and learning...
Principled Technologies
 
PPTX
Internet of Everything -Basic concepts details
sendilkumar66
 
PDF
“A New Era of 3D Sensing: Transforming Industries and Creating Opportunities,...
Edge AI and Vision Alliance
 
PDF
The-2025-Engineering-Revolution-AI-Quality-and-DevOps-Convergence.pdf
Kalilur Rahman
 
PPT
Galois Field Theory of Risk: A Perspective, Protocol, and Mathematical Backgr...
Melanie Swan
 
PDF
Advancing precision in air quality forecasting through machine learning integ...
IAESIJAI
 
PDF
Transform-Your-Factory-with-AI-Driven-Quality-Engineering.pdf
Kalilur Rahman
 
PDF
sbt 2.0: go big (Scala Days 2025 edition)
Eugene Yokota
 
PDF
INTERSPEECH 2025 「Recent Advances and Future Directions in Voice Conversion」
NU_I_TODALAB
 
PDF
The-Future-of-Automotive-Quality-is-Here-AI-Driven-Engineering.pdf
Kalilur Rahman
 
PDF
Statistics on Ai - sourced from AIPRM.pdf
AmelynLeeMeira
 
PDF
Convolutional neural network based encoder-decoder for efficient real-time ob...
IAESIJAI
 
PDF
Consumable AI The What, Why & How for Small Teams.pdf
Vivek Sai
 
PDF
Produktkatalog für HOBO Datenlogger, Wetterstationen, Sensoren, Software und ...
Metrics GmbH
 
PPTX
Microsoft User Copilot Training Slide Deck
zohoron1
 
PDF
AI.gov: A Trojan Horse in the Age of Artificial Intelligence
Michael Weaver
 
PDF
The influence of sentiment analysis in enhancing early warning system model f...
IAESIJAI
 
PDF
5-Ways-AI-is-Revolutionizing-Telecom-Quality-Engineering.pdf
Kalilur Rahman
 
PPTX
future_of_ai_comprehensive_20250822032121.pptx
forrandomuse090
 
Lung cancer patients survival prediction using outlier detection and optimize...
IAESIJAI
 
Dell Pro Micro: Speed customer interactions, patient processing, and learning...
Principled Technologies
 
Internet of Everything -Basic concepts details
sendilkumar66
 
“A New Era of 3D Sensing: Transforming Industries and Creating Opportunities,...
Edge AI and Vision Alliance
 
The-2025-Engineering-Revolution-AI-Quality-and-DevOps-Convergence.pdf
Kalilur Rahman
 
Galois Field Theory of Risk: A Perspective, Protocol, and Mathematical Backgr...
Melanie Swan
 
Advancing precision in air quality forecasting through machine learning integ...
IAESIJAI
 
Transform-Your-Factory-with-AI-Driven-Quality-Engineering.pdf
Kalilur Rahman
 
sbt 2.0: go big (Scala Days 2025 edition)
Eugene Yokota
 
INTERSPEECH 2025 「Recent Advances and Future Directions in Voice Conversion」
NU_I_TODALAB
 
The-Future-of-Automotive-Quality-is-Here-AI-Driven-Engineering.pdf
Kalilur Rahman
 
Statistics on Ai - sourced from AIPRM.pdf
AmelynLeeMeira
 
Convolutional neural network based encoder-decoder for efficient real-time ob...
IAESIJAI
 
Consumable AI The What, Why & How for Small Teams.pdf
Vivek Sai
 
Produktkatalog für HOBO Datenlogger, Wetterstationen, Sensoren, Software und ...
Metrics GmbH
 
Microsoft User Copilot Training Slide Deck
zohoron1
 
AI.gov: A Trojan Horse in the Age of Artificial Intelligence
Michael Weaver
 
The influence of sentiment analysis in enhancing early warning system model f...
IAESIJAI
 
5-Ways-AI-is-Revolutionizing-Telecom-Quality-Engineering.pdf
Kalilur Rahman
 
future_of_ai_comprehensive_20250822032121.pptx
forrandomuse090
 

Instance-Based Ontological Knowledge Acquisition

  • 1. Instance-Based Ontological Knowledge Acquisition The Graduate University for Advanced Studies (SOKENDAI) National Institute of Informatics Lihua Zhao & Ryutaro Ichise ESWC2013, Montpellier, France, 28th May, 2013
  • 2. Outline Introduction Related Work Semi-automatic Ontology Integration Framework Graph-Based Ontology Integration Machine-Learning-Based Ontology Schema Extraction Ontology Merger Experiments Conclusion and Future Work Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 2
  • 3. Introduction Linked Open Data (LOD) Machine-readable and interlinked at instance-level. 295 data sets, 31 billion RDF triples (as of Sep. 2011). Around 504 million owl:sameAs links. 7 domains (cross-domain, geographic, media, life sciences, government, user-generated content, and publications). World Fact- book John Peel (DBTune) Pokedex Pfam US SEC (rdfabout) Linked LCCN Europeana EEA IEEE ChEMBL Semantic XBRL SW Dog Food CORDIS (FUB) AGROVOC Openly Local Discogs (Data Incubator) DBpedia yovisto Tele- graphis tags2con delicious NSF Medi Care Brazilian Poli- ticians dotAC ERA Open Cyc Italian public schools UB Mann- heim JISC Moseley Folk Semantic Tweet OS GTAA totl.net OAI Portu- guese DBpedia LOCAH KEGG Glycan CORDIS (RKB Explorer) UMBEL Affy- metrix riese business. data.gov. uk Open Data Thesau- rus Geo Linked Data UK Post- codes Smart Link ECCO- TCP UniProt (Bio2RDF) SSW Thesau- rus RDF ohloh Freebase London Gazette Open Corpo- rates Airports GEMET P20 TCM Gene DIT Source Code Ecosystem Linked Data OMIM Hellenic FBD Data Gov.ie Music Brainz (DBTune) data.gov.uk intervals LODE Climbing SIDER Project Guten- berg Music Brainz (zitgist) ProDom HGNC SMC Journals Reactome National Radio- activity JP legislation data.gov.uk AEMET Product Types Ontology Linked User Feedback Revyu Gene Ontology NHS (En- AKTing) URI Burner DB Tropes Eurécom ISTAT Immi- gration Lichfield Spen- ding Surge Radio Euro- stat (FUB) Piedmont Accomo- dations New York Times Klapp- stuhl- club EUNIS Bricklink reegle CO2 Emission (En- AKTing) Audio Scrobbler (DBTune) GovTrack GovWILD ECS South- ampton EPrints KEGG Reaction Linked EDGAR (Ontology Central) LIBRIS Open Library KEGG Drug research. data.gov. uk VIVO Cornell UniRef WordNet (RKB Explorer) Cornetto medu- cator DDC Deutsche Bio- graphie Wiki Ulm NASA (Data Incu- bator) BBC Music Drug Bank Turismo de Zaragoza Plymouth Reading Lists education. data.gov. uk KISTI Uni Pathway Eurostat (Ontology Central) OGOLOD Twarql Music Brainz (Data Incubator) Geo Names Pub Chem Italian Museums Good- win Family flickr wrappr Eurostat Thesau- rus W Open Library (Talis) LOIUS Linked GeoData Linked Open Colors WordNet (VUA) patents. data.gov. uk Greek DBpedia Sussex Reading Lists Metoffice Weather Forecasts GND LinkedCT SISVU transport. data.gov. uk Didac- talia dbpedia lite BNB Ontos News Portal LAAS Product DB iServe Recht- spraak. nl KEGG Com- pound Geo Species VIVO UF Linked Sensor Data (Kno.e.sis) lobid Organi- sations LEM Linked Crunch- base FTS Ocean Drilling Codices Janus AMP ntnusc Weather Stations Amster- dam Museum lingvoj Crime (En- AKTing) Course- ware PubMed ACM BBC Wildlife Finder Calames Chronic- ling America data- open- ac- uk Open Election Data Project Slide- share2RDF Finnish Munici- palities OpenEI MARC Codes List VIVO Indiana Hellenic PD LCSH FanHubz bible ontology IdRef Sudoc KEGG Enzyme NTU Resource Lists PRO- SITE Linked Open Numbers Energy (En- AKTing) Roma Open Calais data bnf.fr lobid Resources IRIT theses. fr LOV Rådata nå! Daily Med Taxo- nomy New- castle Google Art wrapper Poké- pédia EURES BibBase RESEX STITCH PDB EARTh IBM Last.FM artists (DBTune) YAGO ECS (RKB Explorer) Event Media STW my Experi- ment BBC Program- mes NDL subjects Taxon Concept Pisa KEGG Pathway UniParc Jamendo (DBtune) Popula- tion (En- AKTing) Geo- WordNet RAMEAU SH UniSTS Mortality (En- AKTing) Alpine Ski Austria DBLP (RKB Explorer) Chem2 Bio2RDF MGI DBLP (L3S) Yahoo! Geo Planet GeneID RDF Book Mashup El Viajero Tourism Uberblic Swedish Open Cultural Heritage GESIS data dcs Last.FM (rdfize) Ren. Energy Genera- tors Sears RAE2001 NSZL Catalog Homolo- Gene Ord- nance Survey TWC LOGD Disea- some EUTC Produc- tions PSH WordNet (W3C) semantic web.org Scotland Geo- graphy Magna- tune Norwe- gian MeSH SGD Traffic Scotland statistics. data.gov. uk Crime Reports UK UniProt US Census (rdfabout) Man- chester Reading Lists EU Insti- tutions PBAC VIAF UN/ LOCODE Lexvo Linked MDB ESD stan- dards reference. data.gov. uk t4gm info Sudoc ECS South- ampton ePrints Classical (DB Tune) DBLP (FU Berlin) Scholaro- meter St. Andrews Resource Lists NVD Fishes of TexasScotland Pupils & Exams RISKS gnoss DEPLOY InterPro Lotico Ox Points Enipedia ndlna Budapest CiteSeer Media Geographic Publications User-generated content Government Cross-domain Life sciences As of September 2011 Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 3
  • 4. Motivation Figure: Interlinked Instances of “France”. Problems when access to several data sets: Ontology Heterogeneity Problem Map related ontology classes and properties. Ontology similarity matching on the SameAs graph patterns. Difficulty in Identifying Core Ontology Schemas Retrieve frequently used core ontology classes and properties. Machine learning for core ontology schema extraction. Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 4
  • 5. Related Work Find useful attributes from frequent graph patterns using a supervised machine learning method. [Le, 2010] Only for geographic data and no discussion about the features. A debugging method for mapping lightweight ontologies with machine learning method. [Meilicke, 2008] Limited to the expressive lightweight ontologies. Construct intermediate-layer ontology by analyzing concept coverings. [Parundekar, 2012] Only for specific domains and limited between two resources. Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 5
  • 6. Semi-automatic Ontology Integration Framework Construct a global ontology by integrating heterogeneous ontologies of the Linked Open Data. Graph-Based Ontology Integration [Zhao, et al., 2012] Group related classes and properties. Machine-Learning-Based Ontology Schema Extraction Extract frequent core classes and properties. Ontology Merger Merge extracted ontology classes and properties. Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 6
  • 7. Semi-automatic Ontology Integration Framework Construct a global ontology by integrating heterogeneous ontologies of the Linked Open Data. Graph-Based Ontology Integration [Zhao, et al., 2012] Group related classes and properties. Machine-Learning-Based Ontology Schema Extraction Extract frequent core classes and properties. Ontology Merger Merge extracted ontology classes and properties. Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 7
  • 8. Graph-based Ontology Integration Extract graph patterns from interlinked instances to discover related ontology classes and predicates. Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 8
  • 9. STEP 1: Graph Pattern Extraction SameAs Graph: An undirected SameAs Graph SG = (V , E, I), where V : a set of vertices (the labels of data sets). E ⊆ V × V : a set of sameAs edges. I: a set of URIs of the interlinked SameAs Instances. Example: SGFrance = (VFrance, EFrance, IFrance). VFrance = {M, D, G, N} EFrance = {(D, G), (D, N), (G, M), (G, N)} IFrance = {mdb-country:FR, db:France, geo:3017382, nyt:67...21}. Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 9
  • 10. STEP 2: <Predicate, Object> Collection <Predicate, Object> (PO) pairs and types for SGFrance Predicate Object Type rdf:type db-onto:Country Class rdfs:label “France”@en String foaf:name “France”@en String foaf:name “R´epublique fran¸caise”@en String db-onto:wikiPageExternalLink http://us.franceguide.com/ URI db-prop:populationEstimate 65447374 Number . . . . . . . . . . . . . . . . . . geo-onto:name France String geo-onto:alternateName “France”@en String geo-onto:featureCode geo-onto:A.PCLI Class geo-onto:population 64768389 Number . . . . . . . . . . . . . . . . . . rdf:type mdb:country Class mdb:country name France String mdb:country population 64094000 Number rdfs:label France (Country) String . . . . . . . . . . . . . . . . . . rdf:type skos:Concept Class skos:inScheme nyt:nytd geo Class skos:prefLabel “France”@en String nyt-prop:first use 2004-09-01 Date Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 10
  • 11. STEP 3: Related Classes and Properties Grouping Related Classes Grouping (Leaf nodes) Tracking subsumption relations from SameAs graphs. < C1 owl:subClassOf C2 > < C1 skos:inScheme C2 > Example: SGFrance Related Classes → {db-onto:Country, geo-onto:A.PCLI, mdb:country, nyt:nytd geo } Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 11
  • 12. STEP 3: Related Classes and Properties Grouping Related Properties Grouping Exact matching for creating initial sets of PO pairs S1, S2, . . . , Sk . Similarity matching on the initial sets of PO pairs. Sim(POi , POj ) = ObjSim(POi , POj ) + PreSim(POi , POj ) 2 ObjSim(POi , POj ) =    1 − |OPOi −OPOj | OPOi +OPOj if OPO is Number StrSim(OPOi , OPOj ) if OPO is String PreSim(POi , POj ) = WNSim(TPOi , TPOj ) StrSim(OPOi , OPOj ): Average of 3 string-based similarity measures. WNSim(TPOi , TPOj ): Average of 9 WordNet-based similarity measures. Refine sets of PO pairs according to rdfs:domain. Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 12
  • 13. STEP 4: Aggregation of All Graph Patterns Aggregate the integrated classes and properties from all the extracted graph patterns. Select A Term for Each Set ex-onto:ClassTerm ex-onto:propTerm Construct Relations ex-prop:hasMemberClasses <class, ex-prop:hasMemberClasses, ex-onto:ClassTerm> ex-prop:hasMemberDataTypes <property, ex-prop:hasMemberDataTypes, ex-onto:propTerm> Construct A Preliminary Integrated Ontology Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 13
  • 14. STEP 5: Manual Revision Manually revise the preliminary integrated ontology. Terms of the integrated classes and properties: Choose a proper term for each group of classes or properties. Groups of related classes or properties: Correct wrong grouping. Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 14
  • 15. Semi-automatic Ontology Integration Framework Construct a global ontology by integrating heterogeneous ontologies of the Linked Open Data. Graph-Based Ontology Integration [Zhao, et al., 2012] Group related classes and properties. Machine-Learning-Based Ontology Schema Extraction Extract frequent core classes and properties. Ontology Merger Merge extracted ontology classes and properties. Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 15
  • 16. Machine-Learning-Based Ontology Schema Extraction Top-level classes and core properties are necessary. Decision Table Retrieves core properties in each data set. Belongs to rule-based machine learning with simple hypothesis. Retrieves a subset of properties that are important for describing instances in a data set. Apriori Retrieves core properties in the instances of a specific top-level class. Belongs to association rule mining. Finds a set of properties, whose support is greater than the user-defined minimum support. Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 16
  • 17. Decision Table Retrieve top-level classes and core properties that are important for describing instances in a data set. Collect top-level classes. Filter out infrequent properties. Convert each instance for the Decision Table algorithm. weight(prop1, inst), weight(prop2, inst), ... weight(propn, inst), class PF-IIF (Property Frequency - Inverse Instance Frequency) weight(prop, inst) = pf (prop, inst) × iif (prop, D) pf (prop, inst) = the frequency of prop in inst. iif (prop, D) = log |D| |instprop| instprop: an instance that contains the property prop. Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 17
  • 18. Apriori Retrieve top-level classes and frequent core properties that are important for describing instances in a specific class. Collect top-level classes. Filter out infrequent properties. Convert each instance of top-level class c for the Apriori algorithm. [prop1, prop2, ..., propn] Define minimum support and confidence metric. Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 18
  • 19. Semi-automatic Ontology Integration Framework Construct a global ontology by integrating heterogeneous ontologies of the Linked Open Data. Graph-Based Ontology Integration [Zhao, et al., 2012] Group related classes and properties. Machine-Learning-Based Ontology Schema Extraction Extract frequent core classes and properties. Ontology Merger Merge extracted ontology classes and properties. Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 19
  • 20. Ontology Merger Graph-Based Ontology Integration outputs a Preliminary Integrated Ontology. For the ontology classes and properties retrieved from Machine-Learning-Based Approach: If Class c ∈ Preliminary Integrated Ontology, add < ex-onto:ClassTermnew , ex-prop:hasMemberClasses, c >. For each Property prop retrieved from top-level class c using Apriori, add a triple < prop, rdfs:domain, c >. Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 20
  • 21. Experiments Data Sets Graph-Based Ontology Integration Decision Table Apriori Comparison of Integrated Ontology Case Studies Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 21
  • 22. Data Sets DBpedia (v3.6): cross-domain, 3.5 million things, 8.9 million URIs. Geonames (v2.2.1): geographical domain, 7 million URIs. NYTimes: media domain, 10,467 subject news. LinkedMDB: media domain, 0.5 million entities. Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 22
  • 23. Data Sets - Machine Learning Data Set Instances Selected Class Top-level Property Selected Instances Class Property DBpedia 3,708,696 64,460 241 28 1385 840 Geonames 7,480,462 45,000 428 9 31 21 NYTimes 10,441 10,441 5 4 8 7 LinkedMDB 694,400 50,000 53 10 107 60 Selected Instances Randomly select instances per class: DBpedia (5000), Geonames(3000), NYTimes(All), LinkedMDB(3000) Top-level Classes Ontology-based data set: Use subsumption relations. Without ontology: Use categories. Selected Properties With frequency threshold θ as √ n, where n is the total number of instances in the data set. Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 23
  • 24. Graph-Based Ontology Integration 13 graph patterns Frequent graph patterns: GP1, GP2, GP3 N,G,D: GP4, GP5, GP7, GP8 N,M,D: GP6 M,G,D: GP9 M,D,N,G: GP10, GP11, GP12, GP13 13 graph patterns. 97 classes into 48 groups. 357 properties into 38 groups. Retrieved related classes and properties by analyzing graph patterns. [Zhao, I-Semantics2012] Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 24
  • 25. Evaluation of Machine Learning Approaches Evaluate the Decision Table and Apriori algorithm. Evaluation of Decision Table Evaluate whether the retrieved sets of properties are important for describing instances by testing if they can be used to distinguish different types of instances in the data set. Evaluation of Apriori Analyze the performance of Apriori algorithm in each data set with examples of retrieved sets of properties. Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 25
  • 26. Decision Table Data Set Precision Recall F-Measure Selected Properties DBpedia 0.892 0.821 0.837 53 Geonames 0.472 0.4 0.324 10 NYTimes 0.795 0.792 0.785 5 LinkedMDB 1 1 1 11 Core properties are evaluated by predicting classes of instances (10-fold). 11 properties from LinkedMDB can correctly identify class of an instance. DBpedia and NYTimes performs good with selected properties. 10 properties from Geonames are commonly used for all types of classes. Examples of retrieved core properties. DBpedia: db-prop:city, db-prop:debut, db-onto:formationYear,etc. Geonames: geo-onto:alternateName, geo-onto:countryCode, etc. NYTimes: nyt:latest use, nyt:topicPage, wgs84 pos:long, etc. LinkedMDB: mdb:director directorid, mdb:writer writerid, etc. Retrieved top-level classes and core properties in each data set. Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 26
  • 27. Apriori Examples of retrieved core properties with Apriori Algorithm. Data Set Class Properties DBpedia db:Event db-onto:place, db-prop:date, db-onto:related/geo. db:Species db-onto:kingdom, db-onto:class, db-onto:family. db:Person foaf:givenName, foaf:surname, db-onto:birthDate. Geonames geo:P geo-onto:alternateName, geo-onto:countryCode. geo:R wgs84 pos:alt, geo-onto:name, geo-onto:countryCode. NYTimes nyt:nytd geo wgs84 pos:long. nyt:nytd des skos:scopeNote. LinkedMDB mdb:actor mdb:performance, mdb:actor name, mdb:actor netflix id. mdb:film mdb:director, mdb:performane, mdb:actor, dc:date. DBpedia and LinkedMDB: Retrieved unique properties. Geonames and NYTimes: Retrieved commonly used properties only. Automatically added missing domain information: < prop, rdfs : domain, classtop >. Retrieved frequent core properties in each top-level class. Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 27
  • 28. Comparison of Integrated Ontology Previous Work Machine Learning Current Work Graph-Based Decision Apriori Integrated Integration Table Ontology Class 97 50 (38 new) 50 (38 new) 135 (38 new) Property 357 79 (49 new) 119(80 new) 453 (96 new) Previous Work: 97 classes in 49 groups, 357 properties in 38 groups. Current Work: 135 classes in 87 groups, 453 properties in 97 groups. Apriori retrieves more properties than Decision Table. 33 new properties are found with both Apriori and Decision Table. Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 28
  • 29. Case Studies I Find Missing Links of Islands with Integrated Ontology SELECT DISTINCT ?geo ?db ?string where { ?geo geo-onto:featureCode geo-onto:T.ISL. ?geo ?gname ?string. ex-onto:name ex-prop:hasMemberDataTypes ?gname. ?db rdf:type db-onto:Island. ex-onto:name ex-prop:hasMemberDataTypes ?dname. ?db ?dname ?string. } Retrieved 509 links, including 218 existing SameAs links: 97 existing links from DBpedia to Geonames. 211 links from Geonames to DBpedia. 90 bidirectional links between DBpedia and Geonames. Discovered 291 missing links with the integrated ontology using exact matching on the labels of instances. Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 29
  • 30. Case Studies II Predicates grouped in ex-prop:birthDate Property Number of Instances rdfs:domain db-onto:birthDate 287,327 db-onto:Person db-prop:datebirth 1,675 N/A db-prop:dateofbirth 87,364 N/A db-prop:dateOfBirth 163,876 N/A db-prop:born 34,832 N/A db-prop:birthdate 70,630 N/A db-prop:birthDate 101,121 N/A Suggest “db-onto:birthDate” as the standard property because it has rdfs:domain definition has the highest usage in the DBpedia instances. Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 30
  • 31. Case Studies III Give me all the cities with more than 10,000,000 inhabitants. Standard Query Query with the Integrated Ontology SELECT DISTINCT ?uri ?string SELECT DISTINCT ?uri ?string WHERE { WHERE { ?uri rdf:type db-onto:City. ?uri rdf:type db-onto:City. ex-onto:population ex-prop:hasMemberDataTypes ?prop. ?uri db-prop:populationTotal ?inhabitants. ?uri ?prop ?inhabitants. FILTER (?inhabitants > 10000000). FILTER (?inhabitants > 10000000). OPTIONAL { ?uri rdfs:label ?string. OPTIONAL { ?uri rdfs:label ?string. FILTER (lang(?string) = ’en’) }} FILTER (lang(?string) = ’en’) }} A SPARQL example from QALD-1 Open Challenge. Standard query: 9 cities. Query with the integrated ontology: 20 cities. Help QA systems for finding more related answers with simple queries. Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 31
  • 32. Conclusion and Future Work Conclusion Semi-automatic ontology integration framework Graph-Based Ontology Integration. Ontology similarity matching on SameAs graph patterns. Retrieve related ontology classes and properties. Machine-Learning-Based Ontology Schema Extraction Decision Table and Apriori. Extract top-level classes and core properties. Ontology Merger Find missing links, detect misuses of ontologies, and access various data sets with integrated ontology. Future Work Automatically detect and revise mistakes in ontology merger. Automatically detect ranges and domains of properties. Test our framework with more LOD data sets. Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 32
  • 33. Thank you! Questions? Lihua Zhao, [email protected] Ryutaro Ichise, [email protected] Lihua Zhao & Ryutaro Ichise | Instance-Based Ontological Knowledge Acquisition | 33