Data Interlinking

Data interlinking
J´erˆome Euzenat
Montbonnot, France
Jerome.Euzenat@inria.fr
June 10, 2015

Data interlinling
Similarity-based approach
Key-based interlinking
Ontology matching & data interlinking
Tools
The problem: RDF data interlinking
3
http://data.bnf.fr/12144801/edgar allan poe the gold bug/, dc:title, “The gold bug”
The gold bug
title
creator
en
E. Poe
lang
firstname lastname
Writer
Work
rdf:type
rdf:type
b a1 a2
Baudelaire Malarmé
The raven
orig
name name
name
orig
authortranslator translator
Person
Book
rdf:type
rdf:type
≈
≥
≤
≥
Jérôme Euzenat Data interlinking 3 / 0

Data interlinling
Tools
Goal of the lecture
Provide an overview of the problem of data interlinking
Describe broad categories of solutions
Point to useful tools for generating links
Mostly about generating links, not on ﬁnding how to generate them

Data interlinling
Tools
Outline
Data interlinling
Tools

Data interlinling
Tools
Data interlinking
I use (with the same meaning):
instance matching
entity linking
data interlinking
I do not use:
record linkage
data deduplication
entity reconciliation
coreference resolution

Data interlinling
Tools
The data interlinking problem
Data interlinking is the task of ﬁnding same entities within diﬀerent datasets
(RDF graphs).
Data source 1 Data source 2
interlinking
owl:sameAs

Data interlinling
Tools
The data interlinking process
Data source
Data source
interlinking Resulting linksSample links
parameters
resources

Data interlinling
Tools
The data interlinking process (2)
d
d
extraction
Linkage spec
generation l
interlinking

Data interlinling
Tools
Approaches to data interlinking
There are two main approaches to data interlinking:
similarity-based: resources are compared through a similarity measure
and if they are similar enough, they are the same.
key-based: suﬃcient conditions for two resources to be the same are
induced and used to ﬁnd same entities

Data interlinling
Tools
Classiﬁcation of similarities
Data interlinking techniques may be based on:
Data ID (URIs);
Data keys
External relations: (explicit or implicit) links to other resources
Data description (content)

Data interlinling
Tools
Manual resource matching
URI1 URI2
Manual observation
owl:sameAs
This does not scale.
But may be good for a ﬁrst sample or reference.
Crowdsourcing?

Data interlinling
Tools
URI matching
URI1 URI2
URI transformation
owl:sameAs
http://dbpedia.org/resource/Johann Sebastian Bach owl:sameAs
http://www.lastfm.fr/music/Johann+Sebastian+Bach
http://rdf.insee.fr/geo/regions-2011.rdf#REG 11 ?
http://ec.europa.eu/eurostat/ramon/rdfdata/nuts2008/FR10

Data interlinling
Tools
Id matching
id id
Finding same ids
owl:sameAs
You can ﬁnd such types of ids:
Social security numbers
ISBN, DOI, MAC addresses, etc.
authorities: ISO (countries, languages), IATA (airports)
Most databases are built on such identiﬁers. . . but they are often local to the
database.

Data interlinling
Tools
Context-based similarity
URI1 URI2
VIAF
Context-based
“similarity”
owl:sameAs
Process:
Project your data into another resource (DBPedia, geonames, viaf, etc.)
Assess relations between considered terms
Import the relation in the dataset

Data interlinling
Tools
Content-based similarity
3
The gold bug
title
creator
E. Poe
ﬁrstname lastname
Writer
Work
rdf:type
rdf:type
b a1 a2
Baudelaire Poe
Le corbeauLe scarab´e d’or
orig
name name
title
authortranslator
Person
Book
rdf:type
rdf:type
Compute similarity
owl:sameAs
Two main approaches:
bag of text
structured similarity

Data interlinling
Tools
Term-based similarity
The gold bug
E. Poe
ﬁrstname lastname
Writer
Work
type
type Baudelaire Poe
Le corbeau
Le scarab´e d’or
orig
name name
title
authortranslator
Person
Book
type type
Compute “bag of words” similarity
owl:sameAs
Various tools:
Normalisation (Stemmer, Tokenizers)
Use of linguistic resources (Wordnet)
Translation
Many similarity measures, especially from information retrieval

Data interlinling
Tools
Structure similarity
title
creator
ﬁrstname lastname
type
type orig
name name
title
authortranslator
type
type
Compute structure similarity
owl:sameAs
Techniques:
Based on graph matching techniques
Can be used to learn weights on properties (but need matching)
Problem: scalability

Data interlinling
Tools
Cross-lingual RDF data interlinking
http://a.org/Mus999 France
Mus´ee du Louvre
nom
lieu
Paris
99,rue de Rivoli
75001
adresse
ville
rue
zip
http://bb.cn/盧浮宮
盧浮宮
法國巴黎
稱號
位於
owl:sameAs ?

Data interlinling
Tools
Similarity-based data interlinking
RESOURCE RESOURCE
SIMILARITY
owl:sameAs ?
Hypothesis: ↑ similarity ↑ probability that it is the same object
DOCUMENT DOCUMENTSIMILARITY
owl:sameAs ?
Yuzhong Qu, Wei Hu, Gong Cheng: Constructing virtual documents for ontology matching. WWW 2006: 23-31.
DOCUMENT(zh) DOCUMENT(en)
DOCUMENT(en)
translation
DOCUMENT(zh)
translationSIMSIM
SIMILARITY
owl:sameAs ?
BabelNet(IDs) BabelNet(IDs)SIMILARITY
owl:sameAs ?

Data interlinling
Tools
General cross-lingual interlinking
framework
1 Virtual
Documents
3 Similarity
Computation
4 Link
Generation
2 Language
Normalization

Data interlinling
Tools
Building virtual documents by levels
http://dbpedia.org/resource/Charles Perrault
Charles Perrault
dbpedia:France
Level 1
France is a sovereign
country in Western Eu-
rope that includes over-
seas regions and territo-
ries. . .
Level 2

Data interlinling
Tools
Machine translation: parameters
1 Virtual
Documents
2.1 Machine
Translation
2.2 NLP
Preprocessing
3 Similarity
Computation
4 Link
Generation
Level 1
Level 2
ZH→EN
Lowercase+Tokenize
+ Filter stop words
+ Stemming (Porter)
+ Bigrams (terms)
TF+cosine
TF*IDF+cosine
Greedy
Hungarian
32 settings have been explored in total

Data interlinling
Tools
Lcase+Tokenization with TF*IDF at
Level 1
0 - 0.11
0.11 - 0.15
0.15 - 0.25
0.25 - 0.35
0.35 - 0.45
0.45 - 1

Data interlinling
Tools
Adding noise

Data interlinling
Tools
BabelNet method: parameters
1 Virtual
Documents
2 Multilingual
KB Mapping
3 Similarity
Computation
4 Link
Generation
Level 1
Level 2
TF+cosine
TF*IDF+cosine
Greedy
Hungarian

Data interlinling
Tools
Database keys
A set of attributes which uniquely identiﬁes elements of a relation
e.g., Book: isbn, People: ﬁstname, lastname, birthplace, birthdate
usually given and used to check integrity
They may be used for identifying same entities across two databases.
But they require alignments.

Data interlinling
Tools
Example of interlinking with keys and
alignments
Are the resources bnf:cb118949856 and bne:XX1721208 the same?
if BNF ontology states foaf:Person owl:hasKey {foaf:name, dc:dates}
and we have the following alignment
foaf:Person
bnf:cb118949856
Albert Camus
07-11-1913
04-01-1960
Romancier, dramaturge et essayiste
http://id.loc.gov/vocabulary/countries/fr
Mondovi (Alg´erie)
1913-1960
foaf:name
rda:dateOfBirth
rda:dateOfDeath
rda:biographicalInformation
rda:countryAssociatedWithThePerson
rda:placeOfBirth
dc:dates
frbrer:C1005
bne:XX1721208
Camus, Albert
1913-1960
Aut [...]1980
frber:P3039
frber:P3040
rda:sourceConsulted
≡
≡
≈
≈
owl:sameAs

Data interlinling
Tools
Key-based interlinking methods
Database keys allow for identifying entities: if they are aligned, this can be
used for linking.
Advantages
they are logically grounded
they allow to minimize the number of properties to compare (if we use
minimal keys)
Drawbacks
Require alignment between properties and classes
Very few key axioms are available, and they are not necessarily useful for
interlinking
We overcome these drawbacks by introducing link keys

Data interlinling
Tools
Link key
A link key
{ p1, q1 , . . . , pn, qn }{ p1, q1 , . . . , pm, qm } linkkey c, d
holds iﬀ
For all pairs of instances a and b belonging respectively to classes c and d of
ontologies O and O ,
if a and b share at least one value (object) for each pairs of
properties pi and qi respectively,
and a and b share all their values (objects) for each pairs of
properties pi and qi respectively,
then they are the same ( a, owl:sameAs, b ).
Example:
{ foaf:name, frbr:P3039 }{ dc:dates, frbr:P3040 } linkkey foaf:Person, frbr:C1005

Data interlinling
Tools
Link key extraction
Problem: How to induce such link keys from data?
The number of set of pairs of properties is exponential
Our approach:
discover only candidate link keys.
evaluate them in order to select only the “good” ones

Data interlinling
Tools
Candidate link key
A candidate link key is a set of property pairs { p1, q1 , . . . , pk, qk } that
1. would generate at least one link if used as a link key
2. is maximal for at least one link, or is the intersection of several
candidate link keys

Data interlinling
Tools
Supervised selection measures
If a sample of reference links is available:
Positive examples (L+) : a set of owl:sameAs links
Negative examples (L−) : a set of owl:diﬀerentFrom links
Idea: Approximate precision and recall on that sample
Deﬁnition (Relative precision and recall)
precision(K, L+
, L−
) =
|L+ ∩ LD,D (K)|
|(L+ ∪ L−) ∩ LD,D (K)|
recall(K, L+
) =
|L+ ∩ LD,D (K)|
|L+|

Data interlinling
Tools
Unsupervised selection measures
When no reference link is available.
Idea: measuring how close the extracted links would be from
one-to-one and total.
Deﬁnition (Discriminability)
disc(K, D, D ) =
min(|{a : a, b ∈ LD,D (K)}|, |{b : a, b ∈ LD,D (K)}|)
|LD,D (K)|
Deﬁnition (Coverage)
cov(K, D, D ) =
|{a : a, b ∈ LD,D (K)} ∪ {b : a, b ∈ LD,D (K)}|
|{a : c(a) ∈ D} ∪ {b : d(b) ∈ D }|

Data interlinling
Tools
Experimental evaluation
These selection measures were evaluated on public datasets.
Finding links between French municipalities described in two diﬀerent
datasets:
Insee dataset: 36700 instances;
Geonames dataset: 36552 instances.
The reference link set is composed of:
Positive links: 36552 owl:sameAs statements;
owl:diﬀerentFrom links derived from owl:sameAs links (closed world
assumption).

Data interlinling
Tools
Evaluation
The algorithm extracted 11 candidate link keys:
{1} {2} {3, 4} {5, 6}
{7, 1} {2, 1} {3, 4, 1} {3, 2, 4}
{3, 7, 4, 1} {3, 2, 4, 1}
{3, 7, 2, 4, 1}
coverage
discriminability
1 = nom, name 2 = nom, alternateName
3 = subdivisionDe, parentFeature 4 = subdivisionDe, parentADM3
5 = codeINSEE, population 6 = codeCommune, population
7 = nom, oﬃcialName

Data interlinling
Tools
Evaluation
Correlation between the harmonic means of discriminability and coverage and
F-measure:
bad F-measure≈ 0
high F-measure≈ .99
good F-measure≈ 0.89
{1} {2} {3, 4} {5, 6}
{7, 1} {2, 1} {3, 4, 1} {3, 2, 4}
{3, 7, 4, 1} {3, 2, 4, 1}
{3, 7, 2, 4, 1}
h-mean(disc.,cov)≈ .99 h-mean(disc.,cov)≈ .89 h-mean(disc.,cov) ≈ 0
1 = nom, name 2 = nom, alternateName
3 = subdivisionDe, parentFeature 4 = subdivisionDe, parentADM3
5 = codeINSEE, population 6 = codeCommune, population
7 = nom, oﬃcialName

Data interlinling
Tools
Why using ontologies?
Because it is obvious that we must compare the instances of equivalent
classes based on equivalent properties.
More precisely:
For reducing the search space for ﬁnding link keys and similarities
For reducing the scope of linkage speciﬁcations
Because not the same linkage rules work for the same classes
Because classes and properties are hint like others of the similarity
between resources
Ex. With similarity and with keys

Data interlinling
Tools
Data interlinking through a common
ontology
o
URI1 URI2
Resource matching
of datasets
described by the
same ontology
owl:sameAs

Data interlinling
Tools
Matching with a common ontology
+ Focus the search: only match instances of the same class;
– Not suﬃcient: it remains to identify corresponding entities
+ If keys are deﬁned (OWL 2), this is done;
+ At least we know which properties to compare;
– Inferring secondary keys may be useful;
– Correcting discrepancies: record linkage.

Data interlinling
Tools
Record linkage
Name Johann
Date 1665-03-21
Place M¨unchen
NameJohannes
Date31/03/1665
PlaceMonaco di Bavaria
Having a common ontology does not solve all problems.

Data interlinling
Tools
Different types of mismatch
Different domains, connected (BIM, Energy demand)
⇒ few correspondences, any type
Same domain, different models (engineer, policy maker)
⇒ many correspondences, mostly equivalence
Same domain, different granularity (city management, building design)
⇒ many correspondences, mostly subsumption

Data interlinling
Tools
Data interlinking with diﬀerent
ontologies (implicit alignment)
o o
URI1 URI2
Resource matching
of datasets
described by
diﬀerent ontologies
owl:sameAs

Data interlinling
Tools
Data interlinking with diﬀerent
ontologies (explicit alignment)
o o
URI1 URI2
A
Resource matching
of datasets
described by
diﬀerent ontologies
owl:sameAs

Data interlinling
Tools
Ontology matching for data interlinking
o o
URI1 URI2
Ontology matching
A
Data interlinking
owl:sameAs

Data interlinling
Tools
Heterogeneity problem
Resources being expressed in different ways must be reconciled before being
used.
Mismatch between formalized knowledge can occur when:
different languages are used (OWL vs. Topic maps);
different terminologies are used:
English vs. Chinese;
Book vs. Monograph.
different models are used:
different classes: Autobiography vs. Paperback;
classes vs. property: Essay vs. literarygenre;
classes vs. instances: One physical book as an instance vs. one work as
an instance.
different scopes and granularity are used.
Only books vs. cultural items vs. any product;
Books detailed to the print and translation level vs. books as works.

Data interlinling
Tools
Ontology alignment
Item
DVD
Book
Paperback
Hardcover
CD
price
title
doi
creator
pp
author
integer
string
uri
Person
Monograph
Essay
Literary critics
Politics
Biography
Autobiography
Literature
pages
isbn
author
title
subject
Human
Writer
≥
≥
≥
≤
≥

Data interlinling
Tools
Expressive alignments (EDOAL)
Pocket
Book
topic
author
=
Volume
size14
≥
Autobiography
=
∀x, Pocket(x) ⇐ Volume(x) ∧ size(x, y) ∧ y ≤ 14
∀x, Book(x) ∧ author(x, y) ∧ topic(x, y) ≡ Autobiography(x)

Data interlinling
Tools
Example: INSEE dataset
Région table:
code nom chef-lieu
11 Île-de-France 75056
21 Champagne-Ardenne 51108
22 Picardie 80021
Sous-région table:
région département
11 75
11 77
11 78
11 91
11 92
11 93

Data interlinling
Tools
Example: Administrative ontology
Territoire FR
Pays
Region
Departement
Arrondissement
Commune
code
nom
chef-lieu
subdivision
integer
string

Data interlinling
Tools
Example: NUTS dataset
NUTSRegion table:
level code name hasParentRegion
0 FR FRANCE
1 FR1 ˆILE DE FRANCE FR
2 FR10 ˆIle de France FR1
3 FR101 Paris FR10
3 FR104 Essonne FR10

Data interlinling
Tools
Example: Linking INSEE and NUTS
NUTS: Nomenclature of territorial units for statistics
#INSEE INSEE name NUTS Level #NUTS
1 Pays 0 34
1 142
26 R´egion 2 344
100 D´epartement 3 1488
342 Arrondissement
4036 Canton 4
52422 Commune 5

Data interlinling
Tools
Example: Linking INSEE and NUTS
Territoire FR
Pays
Region
Departement
Commune
PAYS FR
REG 11
DEP 75
DEP 77
DEP 78
COM 75056
Region
Country
NUTSRegion
LAURegion
FR
UK
FR1
FR10
FR101
FR102
FR103
owl:sameAs
owl:sameAs
owl:sameAs
owl:sameAs
owl:sameAs

Data interlinling
Tools
Example: Linksets
Speciﬁc data sets containing URIs.
<http://www.example.org/linkset/INSEE-NUTS>
a void:Linkset ;
void:target <http://rdf.insee.fr/geo/regions-2011.rdf>;
void:target <http://nuts.psi.enakting.org/id/>;
insee:PAYS FR owl:sameAs nuts:FR
insee:REG 11 owl:sameAs nuts:FR10
insee:DEP 75 owl:sameAs nuts:FR101

Data interlinling
Tools
Example: interesting sets
nuts
onsordnance s. igninsee
geonames dbpedia freebase

Data interlinling
Tools
A simple algorithm
Find matching concepts [concept matching];
For each of them, determine matching properties based on the similarity
between their values in both datasets [property matching];
From them ﬁnd property combinations identifying corresponding entities
[key extraction];
Link corresponding entities [link generation].
For instance, nom/RegionINSEE ⊆ name/NUTSRegionNUTS and moreover
they are unambiguous.

Data interlinling
Tools
INSEE and NUTS: ontology alignment
Territoire FR
Pays
Region
Departement
Arrondissement
Canton
Commune
code
nom
chef-lieu
subdivision
integer
string
Region
Country
NUTSRegion
LAURegion
name
level
code
hasSubRegion
=
≤
≤
≤
≤
≤
=

Data interlinling
Tools
Simple alignments are not suﬃcient
Territoire FR
Region
Departement
Commune
nom
DEP 75
nom
COM 75056
nom
Region
NUTSRegion
name
FR101
name
Paris
=
=
=
≤
≤
≤
=
=
=

Data interlinling
Tools
Expressive alignments are necessary
Region
NUTSRegion
level
hasParentRegion
2 =
FR
=
=
subdivision hasSubRegion
=
nom name
=

Data interlinling
Tools
What does this mean?
Ontology alignments are schema-level expression of correspondences;
They are useful for focussing the search;
Expressive alignments are necessary;
They can be turned into SPARQL-based link generators.
but it is also necessary to express instance level constraints:
for converting data (e.g., mph vs. m/s);
for expressing matching constraint on data (e.g., similarity).

Data interlinling
Tools
Data interlinking and ontology matching
d
o
d
oMatcher
A
Generator
l

Data interlinling
Tools
Tools for data interlinking
Linkage spec extraction generation
similarity LIMES Silk, LIMES, OpenReﬁne
key LinkKeyDisco SPARQL

Data interlinling
Tools
Silk
Silk is a robust software for interlinking data sets.
It relies on an expressive speciﬁcation of linking conditions:
Declare data sources (DataSource);
Circumscribe entities to compare (Source/TargetDataset);
Describe how to compare them (LinkageRule):
Select properties to compare through paths (Input);
Compute distances between them (Compare+threshold);
Aggregate all comparisons (Aggregate);
Select those pairs of entities to be linked (Filter);
Generate links (Output+thresholds).

Data interlinling
Tools
A Silk script
Consider a linking script between INSEE and NUTS:
<Silk>
<Prefix id="nuts"
namespace="http://ec.europa.eu/.../geographic.rdf#" />
<Prefix id="insee"
namespace="http://rdf.insee.fr/geo/" />
<DataSource id="nuts2008"
type="sparqlEndpoint">
<Param name="endpointURI"
value="http://localhost:9091/.../internal"/>
<Param name="graph"
value="http://localhost:9091/.../nuts2008-complete-1"/>
</DataSource>
<DataSource id="insee2010"
type="sparqlEndpoint">
<Param name="endpointURI"
value="http://localhost:9091/.../internal"/>
<Param name="graph"
value="http://localhost:9091/.../source/regions-2010-1"/>
</DataSource>
<Thresholds accept="0.9" verify="0.7" />
<Outputs>
<Output type="sparul">
<Param name="graphUri"
value="http://localhost:9091/.../source/insee-nuts-silk"/>
<Param name="uri"
value="http://localhost:9091/.../lifted/"/>
<Param name="parameter" value="update"/>
</Output>
</Outputs>
<Interlinks>
<Interlink id="linkingNUTS">
<LinkType>owl:sameAs</LinkType>
<SourceDataset dataSource="nuts2008" var="s">
<RestrictTo>?s rdf:type nuts:NUTSRegion.
?s nuts:level 2.
</RestrictTo>
</SourceDataset>
<TargetDataset dataSource="insee2010" var="ss">
<RestrictTo>?ss rdf:type insee:Region</RestrictTo
</TargetDataset>
<LinkageRule>
<Aggregate type="max">
<Compare metric="levenshteinDistance"
threshold=".2">
<Input path="?s/nuts:name"/>
<Input path="?ss/insee:nom"/>
</Compare>
</Aggregate>
</LinkageRule>
</Interlinks>
</Interlink>
</Silk>

Data interlinling
Tools
Silk: preﬁx and sources
<Silk>
<Prefix id="nuts" namespace="http://ec.europa.eu/.../geographic.rdf#" />
<Prefix id="insee" namespace="http://rdf.insee.fr/geo/" />
<DataSource id="nuts2008" type="sparqlEndpoint">
<Param name="endpointURI" value="http://localhost:9091/.../internal"/>
<Param name="graph" value="http://localhost:9091/.../nuts2008-complete-1"/
</DataSource>
<DataSource id="id1" type="file">
<Param name="file" value="/Skratch/TutoLinking/admin/regions-2010.rdf"/>
<Param name="format" value="RDF/XML" />
</DataSource>
Sources can be ﬁles or SPARQL endpoint.

Data interlinling
Tools
Silk rules
<Interlinks>
<Interlink id="linkingNUTS">
<LinkType>owl:sameAs</LinkType>
<SourceDataset dataSource="nuts2008" var="s">
<RestrictTo>?s rdf:type nuts:NUTSRegion.
?s nuts:level 2.
</RestrictTo>
</SourceDataset>
<TargetDataset dataSource="insee2010" var="ss">
<RestrictTo>?ss rdf:type insee:Region</RestrictTo>
</TargetDataset>
<Thresholds accept="0.9" verify="0.7" />
<Outputs>
<Output type="sparul">
<Param name="graphUri" value="http://localhost:9091/.../source/insee-nut
<Param name="uri" value="http://localhost:9091/.../lifted/"/>
<Param name="parameter" value="update"/>
</Output>
</Outputs>

Data interlinling
Tools
Silk rules (cont’ed)
<LinkageRule>
<Aggregate type="max">
<Compare metric="levenshteinDistance" threshold=".2">
<Input path="?s/nuts:name"/>
<Input path="?ss/insee:nom"/>
</Compare>
</Aggregate>
</LinkageRule>
</Interlinks>
</Interlink>
</Silk>
They can:
transform the data (lowercase, tokenize, to integers, etc.),
use comparison metrics (equality, levenshtein, Jaro-Winkler, etc.), and
aggregate their values (average, min, max, etc.).

Data interlinling
Tools
Silk workbench

Data interlinling
Tools
EDOAL Alignments
<Cell>
<entity1><e:Class rdf:about="&insee;Region"/></entity1>
<entity2>
<e:Class>
<e:and rdf:parseType="Collection">
<e:Class rdf:about="&nuts;NUTSRegion"/>
<e:AttributeValueRestriction>
<e:onAttribute><e:Property rdf:about="&nuts;level"/></e:onAttribute>
<e:comparator rdf:resource="&edoal;equals"/>
<e:value><e:Literal e:type="&xsd;integer" e:string="2" /></e:value>
</e:AttributeValueRestriction>
<e:AttributeValueRestriction>
<e:onAttribute>
<e:Relation rdf:about="&nuts;hasParentRegion" />
</e:onAttribute>
<e:comparator rdf:resource="&edoal;equals"/>
<e:value><e:Instance rdf:about="&esdata;FR" /></e:value>
</e:AttributeValueRestriction>
</e:and>
</e:Class>
</entity2>
<relation>equivalence</relation>J´erˆome Euzenat Data interlinking 72 / 0

Data interlinling
Tools
Link keys in the Alignment API
<e:linkkey>
<e:Linkkey>
<e:binding>
<e:Intersects>
<e:property1><e:Property rdf:about="&insee;nom" /></e:property1>
<e:property2><e:Property rdf:about="&nuts;name" /></e:property2>
</e:Intersects>
<e:Equals>
<e:property1>
<e:Property>
<e:inverse><e:Property rdf:about="&insee;subdivision" /></e:inverse>
</e:property1>
<e:property2><e:Property rdf:about="&nuts;hasParentRegion" /></e:propert
</e:Equals>
</e:binding>
</e:Linkkey>
</e:linkkey>

Data interlinling
Tools
Query generation
PREFIX insee: <http://rdf.insee.fr/ontologie-geo-2006.rdf#>
PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
SELECT ?r
FROM <http://rdf.insee.fr/geo/regions-2011.rdf>
WHERE {
?r rdf:type insee:Region .
}
PREFIX nuts: <http://ec.europa.eu/eurostat/ramon/ontologies/geographi
SELECT ?n
FROM <http://ec.europa.eu/eurostat/ramon/rdfdata/nuts2008/>
WHERE {
?n rdf:type nuts:NUTSRegion .
?n nuts:level 2^^xsd:integer .
?n nuts:hasParentRegion nuts:FR .
}

Data interlinling
Tools
Data transformation
CONSTRUCT {
?r rdf:type nuts:NUTSRegion .
?r nuts:level 2^^xsd:integer .
?r nuts:hasParentRegion nuts:FR .
}
WHERE {
}

Data interlinling
Tools
SameAs link generation generation
PREFIX nuts: <http://ec.europa.eu/eurostat/ramon/ontologies/geographi
CONSTRUCT { ?r owl:sameAs ?n . }
FROM <http://ec.europa.eu/eurostat/ramon/rdfdata/nuts2008/>
WHERE {
?r insee:nom ?l .
?n rdf:type nuts:NUTSRegion .
?n nuts:name ?l .
?n nuts:level 2^^xsd:integer .
?n nuts:hasParentRegion nuts:FR .
}

Data interlinling
Tools
Other issue: performances
n × m n
3 × m
3 + n
3 × m
3 + n
3 × m
3
10 × 10 = 100
1000 × 1000 = 1000000
100000 × 100000 = 10000000000

Data interlinling
Tools
Blocking: index+cluster
Dataset 1 Dataset 2

Data interlinling
Tools
Blocks can be obtained from:
clustering values in index
predeﬁned block (based on equality)
classes in an ontology (blocks are deﬁned as class expressions)

Data interlinling
Tools
Other issue: evaluation
d
d
interlinking l
Reference links
evaluation
Precision
Recall
F-measure

Data interlinling
Tools
Other issue: learning
d
d
Training links interlinking l
evaluation

Data interlinling
Tools
Conclusion
Data interlinking is one of the most critical task in linked data
. . . but not only, e.g. smart cities
If faces many problems due to:
heterogeneity (format, languages, convention)
size
Interlinking can be based on similarities or keys
There is active work to infer such interlinking pattern

Data interlinling
Tools
Further reading
T. Heath, C. Bizer, Linked Data: Evolving the Web into a Global Data
Space, Morgan & Claypool (US), 2011 http://linkeddatabook.com/
J. Euzenat, P. Shvaiko, Ontology matching, 2nd ed., Springer,
Heildelberg (DE), 2013 http://book.ontologymatching.org
K. Stefanidis, V. Efthymiou, M. Herschel, V. Christophides, Entity
Resolution in the Web of Data, Tutorial, WWW conference, Seoul
(KR), 2014 http://www.csd.uoc.gr/~vefthym/er/
Silk http://silk-framework.com/
Alignment API http://alignapi.gforge.inria.fr
Al 4 SC http://al4sc.inrialpes.fr

Data interlinling
Tools
Thanks
To my colleagues Manuel Atencia, Jérôme David, Nicolas Guillouet and
Fran¸cois Scharffe
The Datalift and Lindicle projects
The Ready4SmartCities project

http://exmo.inria.fr
Jerome . Euzenat @ inria . fr

Data Interlinking

More Related Content

What's hot (20)

Viewers also liked (20)

Similar to Data Interlinking (20)

More from LD4SC (8)

Recently uploaded (20)

Data Interlinking