Top Banner
Fixing the Domain and Range of Properties in Linked Data by Context Disambiguation Alberto Tonon, Michele Catasta, Gianluca Demartini, Philippe Cudré-Mauroux LDOW - May the 19th, 2015
18

Fixing the Domain and Range of Properties in Linked Data by Context Disambiguation

Aug 05, 2015

Download

Science

Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Page 1: Fixing the Domain and Range of Properties in Linked Data by Context Disambiguation

Fixing the Domain and Range of Properties in Linked Data by Context Disambiguation

Alberto Tonon, Michele Catasta, Gianluca Demartini, Philippe Cudré-Mauroux

LDOW - May the 19th, 2015

Page 2: Fixing the Domain and Range of Properties in Linked Data by Context Disambiguation

Linked Data…

2

"Cobie Smulders"

"Neil Patrick Harris"

"How I Met Your Mother"

showName

starring

starring

name

name

TV Show

type type

Persontype

type

type

type

network

typeTV Network

Broadcast Network

type

Actor

Actor

Person

Work

Page 3: Fixing the Domain and Range of Properties in Linked Data by Context Disambiguation

… and its Schema

3

......

Thing

Person Work

TV showActor

Organisation

Broadcaster...

Type Hierarchy

network

Broadcaster

range

domain

Broadcasterstarring

Work

range

domain

Actor

Property Definitions

Page 4: Fixing the Domain and Range of Properties in Linked Data by Context Disambiguation

Data-Schema Coherence

4

"Cobie Smulders"

"Neil Patrick Harris"

"How I Met Your Mother"

showName

starring

starring

name

name

TV Show

type type

Persontype

type

type

type

network

typeTV Network

Broadcast Network

type

Actor

Actor

Person

Work

network

Broadcaster

range

domain

Broadcasterstarring

Work

range

domain

Actor

Page 5: Fixing the Domain and Range of Properties in Linked Data by Context Disambiguation

Data-Schema Coherence

4

"Cobie Smulders"

"Neil Patrick Harris"

"How I Met Your Mother"

showName

starring

starring

name

name

TV Show

type type

Persontype

type

type

type

network

typeTV Network

Broadcast Network

type

Actor

Actor

Person

Work

network

Broadcaster

range

domain

Broadcasterstarring

Work

range

domain

Actor

✔✔

Page 6: Fixing the Domain and Range of Properties in Linked Data by Context Disambiguation

Data-Schema Coherence

4

"Cobie Smulders"

"Neil Patrick Harris"

"How I Met Your Mother"

showName

starring

starring

name

name

TV Show

type type

Persontype

type

type

type

network

typeTV Network

Broadcast Network

type

Actor

Actor

Person

Work

network

Broadcaster

range

domain

Broadcasterstarring

Work

range

domain

Actor

✔✔✘

Page 7: Fixing the Domain and Range of Properties in Linked Data by Context Disambiguation

Incoherences in Real KBs

5

Property Dom Incoherences

% Dom Incoherences

dpo:years ~641k 100%dpo:currentMember ~260k 100%… … …

Property Dom Incoherences

% Dom Incoherences

fb:[…]object.type ~99M 61%fb:[…]object.name ~41M 100%… … …

Page 8: Fixing the Domain and Range of Properties in Linked Data by Context Disambiguation

Data-Driven Domains/Ranges

• Just intersect the types of all resources appearing as subject/object…

• …being consistent with the type hierarchy.

6

......

Thing

Person Work

TV showActor

Organisation

Broadcaster...

Type Hierarchy

Page 9: Fixing the Domain and Range of Properties in Linked Data by Context Disambiguation

Data-Driven Domains/Ranges

• Dom(foaf:name) = Thing —> Everything has a name !

• Dom(dpo:manager) = Thing —> Everything has a manager "

7

Page 10: Fixing the Domain and Range of Properties in Linked Data by Context Disambiguation

SportSeason0.55

Agent0.44

...

Thing1.00

...Soccer Cricket"k�1

Rugby"k

Baseball"10.42

... ...SoccerClubSeason0.55

SportsTeam0.44

... ... Organisation0.44

SportsTeamSeason0.55

LEXT: an ExampleComputing the domain of dpo:manager

8

Page 11: Fixing the Domain and Range of Properties in Linked Data by Context Disambiguation

SportSeason0.55

Agent0.44

...

Thing1.00

...Soccer Cricket"k�1

Rugby"k

Baseball"10.42

... ...SoccerClubSeason0.55

SportsTeam0.44

... ... Organisation0.44

SportsTeamSeason0.55

dpo:manager is used in two different contexts

LEXT: an ExampleComputing the domain of dpo:manager

8

Page 12: Fixing the Domain and Range of Properties in Linked Data by Context Disambiguation

SportSeason0.55

Agent0.44

...

Thing1.00

...Soccer Cricket"k�1

Rugby"k

Baseball"10.42

... ...SoccerClubSeason0.55

SportsTeam0.44

... ... Organisation0.44

SportsTeamSeason0.55

dpo:manager is used in two different contexts

LEXT: an ExampleComputing the domain of dpo:manager

8

manager

soccer club season manager

sports teammanager

Thing

SoccerClubSeason SportsTeam

Page 13: Fixing the Domain and Range of Properties in Linked Data by Context Disambiguation

SportSeason0.55

Agent0.44

...

Thing1.00

...Soccer Cricket"k�1

Rugby"k

Baseball"10.42

... ...SoccerClubSeason0.55

SportsTeam0.44

... ... Organisation0.44

SportsTeamSeason0.55

dpo:manager is used in two different contexts

LEXT: an ExampleComputing the domain of dpo:manager

8

manager

soccer club season manager

sports teammanager

Thing

SoccerClubSeason SportsTeam

Visit the hierarchy until: 1) Pr(type | property) ≥ λ&&2) H(Pr(property | children)) < η

LEXT

Page 14: Fixing the Domain and Range of Properties in Linked Data by Context Disambiguation

H = 1.96

H = 0.9 SportSeason0.55

Agent0.44

...

Thing1.00

...Soccer Cricket"k�1

Rugby"k

Baseball"10.42

... ...SoccerClubSeason0.55

SportsTeam0.44

... ... Organisation0.44

SportsTeamSeason0.55

dpo:manager is used in two different contexts

LEXT: an ExampleComputing the domain of dpo:manager

8

manager

soccer club season manager

sports teammanager

Thing

SoccerClubSeason SportsTeam

Visit the hierarchy until: 1) Pr(type | property) ≥ λ&&2) H(Pr(property | children)) < η

LEXT

Page 15: Fixing the Domain and Range of Properties in Linked Data by Context Disambiguation

REXT and LERIXT• REXT = LEXT but with types of object resources

• LERIXT = LEXT + REXT

• two type trees (one for Domain and one for Range), current state is a pair (subject type, object type)

9

SportSeason Agent ...

Thing

...Soccer Cricket RugbyBaseball

... ...SoccerClubSeason SportsTeam

... ... OrganisationSportsTeamSeason

SportSeason Agent ...

Thing

...Soccer Cricket RugbyBaseball

... ...SoccerClubSeason SportsTeam

... ... OrganisationSportsTeamSeason

Current State

Page 16: Fixing the Domain and Range of Properties in Linked Data by Context Disambiguation

About λ

10

Figure 1: Coverage and number of new sub-properties varying λ.

Page 17: Fixing the Domain and Range of Properties in Linked Data by Context Disambiguation

Evaluation• Fixed λ = 0.1, η = 1

• 3 authors + 2 experts (majority vote) evaluated the output of LEXT REXT, and LERIXT.

• LERIXT generates too many new sub-properties

11

LEXT REXT LERIXT

Precision 96.50% 91.40% 87.00%

Table 2: Precision of LEXT, REXT, and LERIXT

Page 18: Fixing the Domain and Range of Properties in Linked Data by Context Disambiguation

Conclusion• Three different methods for identifying contexts

• LEXT: exploits the type of the subject resources

• REXT: exploits the type of the object resources

• LERIXT: exploits both

• Up to 96.50% precision.

12

Visit the hierarchy until: 1) Pr(type | property) ≥ λ&&2) H(Pr(property | children)) < η

LEXT