+ All Categories
Home > Technology > Statistical Schema Induction

Statistical Schema Induction

Date post: 26-Jun-2015
Category:
Upload: johanna-voelker
View: 350 times
Download: 0 times
Share this document with a friend
Description:
presentation at the Extended Semantic Web Conference (ESWC) 2011
Popular Tags:
31
% "! " { johanna, mathias }@informatik.uni-mannheim.de " "! " ""!" #"
Transcript
Page 1: Statistical Schema Induction

{ johanna, mathias }@informatik.uni-mannheim.de

Page 2: Statistical Schema Induction

!""#$%%&'(')*+,-."'&(/)'0-.))!1'-(2134343

�•

�•

�–�–�–

�•

Page 3: Statistical Schema Induction

!""#$%%&'(')*+,-."'&(/)'0-.))!1'-(2134343

Page 4: Statistical Schema Induction

!""#$%%&'(')*+,-."'&(/)'0-.))!1'-(2134343

�•

�•�–�–

�–

�•

Page 5: Statistical Schema Induction

!""#$%%&'(')*+,-."'&(/)'0-.))!1'-(2134343

!r." # C

r1 $ r2 # r3

C1 % C2 # C3

C1 # !r.C2

a2 & C1, C2, C5

a1 & C1, C4

a3 & C2, C3

Page 6: Statistical Schema Induction

!""#$%%&'(')*+,-."'&(/)'0-.))!1'-(2134343

239

7,605

532

532

532

532

1,477,796

2,289,540

C!r.C!r."!r!1."

r $ rr

a(a, b)

SELECT distinct ?x WHERE { ?y a ?x . }

SELECT distinct ?x WHERE { ?y ?x ?z . ?z a ?zt . }

Page 7: Statistical Schema Induction

!""#$%%&'(')*+,-."'&(/)'0-.))!1'-(2134343

Page 8: Statistical Schema Induction

!""#$%%&'(')*+,-."'&(/)'0-.))!1'-(2134343

D = (t1, t2, ..., tm) ti ' I

I = {i1, i2, ..., in}

X & YX, Y ' I X ( Y = )

{ pizza, beer } & { chips, ice cream }

Page 9: Statistical Schema Induction

!""#$%%&'(')*+,-."'&(/)'0-.))!1'-(2134343

supp(X) =| {ti * D : X ' ti} |

| D |

conf(X & Y ) =supp(X + Y )

supp(X)

supp(X & Y ) = supp(X + Y )

Page 10: Statistical Schema Induction

!""#$%%&'(')*+,-."'&(/)'0-.))!1'-(213453453

1 A C D E 2 A B C 3 B C D E 4 B C E

supp(r1) = 0.5

r1: { C, D } & { E }r2: { B } & { E }

supp(r2) = 0.5

conf(r1) = 1.0 conf(r2) =23

Page 11: Statistical Schema Induction

!""#$%%&'(')*+,-."'&(/)'0-.))!1'-(213443443

1146 6330 6973 64 67 3131 3647 3255 3664 3109 3641 64 72 3235 6668 6769 3907 2 66 1088 3739 3961 6271 4102 5505 76 101 1146 1346 6330 25 26 64 68 1146 3788 3993 6330 64 147 148

64 = Person 67 = OfficeHolder 6973 = birthPlace some Country 6330 = birthPlace some Place 1146 = birthPlace some PopulatedPlace

!birthPlace.PopulatedPlace # Person{ 1146 } & { 64 }

Page 12: Statistical Schema Induction

!""#$%%&'(')*+,-."'&(/)'0-.))!1'-(213453453

C # DC % D # ED # !r.C!r.C # D!r." # C!r!1." # Cr # sr $ r # r

{Ci} & {Cj}{Ci, Cj} & {Ck}{Ci} & {!rj .Cjk}{!rj .Cjk} & {Ci}{!rj ."} & {Ci}{!r!1

j ."} & {Ci}{ri} & {rj}{ri $ ri} & {ri}

!birthPlace.PopulatedPlace # Person{ 1146 } & { 64 }

Page 13: Statistical Schema Induction

!""#$%%&'(')*+,-."'&(/)'0-.))!1'-(213453453

EL++

",C % D

!r.C

C # Dr1 $ ... $ rk # r

!I

)CI ( DI

{x * !I | !y * !I : (x, y) * rI - y * CI}CI ' DI

rI1 $ ... $ rIk ' rI

Page 14: Statistical Schema Induction

!""#$%%&'(')*+,-."'&(/)'0-.))!1'-(213453453

Page 15: Statistical Schema Induction

!""#$%%&'(')*+,-."'&(/)'0-.))!1'-(213453453

�•�•

�–

�–

�•�–�–

C # D !r." # C !r!1." # C

Page 16: Statistical Schema Induction

!""#$%%&'(')*+,-."'&(/)'0-.))!1'-(213453453

ALF(D)

259

257

602

674

459

482

http://wiki.dbpedia.org/Ontology

Page 17: Statistical Schema Induction

!""#$%%&'(')*+,-."'&(/)'0-.))!1'-(213453453

rdf:type

C # D

Page 18: Statistical Schema Induction

!""#$%%&'(')*+,-."'&(/)'0-.))!1'-(213453453

!r." # C

Page 19: Statistical Schema Induction

!""#$%%&'(')*+,-."'&(/)'0-.))!1'-(213453453

!r!1." # C

Page 20: Statistical Schema Induction

!""#$%%&'(')*+,-."'&(/)'0-.))!1'-(213453453

�•

�•

<owl:ObjectProperty rdf:about="http://dbpedia.org/ontology/partOf"> <rdfs:label xml:lang="en">part of</rdfs:label> <rdfs:domain rdf:resource="http://dbpedia.org/ontology/MilitaryConflict"/> <rdfs:range rdf:resource="http://dbpedia.org/ontology/MilitaryConflict"/> </owl:ObjectProperty>

!gameEngine!1." # Software!recordLabel." # MusicalWork

Page 21: Statistical Schema Induction

!""#$%%&'(')*+,-."'&(/)'0-.))!1'-(213453453

<owl:Restriction> <rdfs:subClassOf rdf:resource="http://dbpedia.org/ontology/Place"/> <owl:onProperty rdf:resource="http://dbpedia.org/ontology/majorIsland"/> <owl:someValuesFrom rdf:resource="http://dbpedia.org/ontology/Mountain"/> </owl:Restriction>

SELECT ?x ?y WHERE { ?x <http://dbpedia.org/ontology/majorIsland> ?y . ?y a <http://dbpedia.org/ontology/Mountain> . }

Socorro_Island majorIsland San_Benedicto_Island San_Benedicto_Island rdf:type Mountain

!majorIsland.Mountain # Place

Page 22: Statistical Schema Induction

!""#$%%&'(')*+,-."'&(/)'0-.))!1'-(213443443

<owl:Restriction> <rdfs:subClassOf rdf:resource="http://dbpedia.org/ontology/Person"/> <owl:onProperty rdf:resource="http://dbpedia.org/ontology/nationality"/> <owl:someValuesFrom rdf:resource="http://dbpedia.org/ontology/Book"/> </owl:Restriction>

Gerard_David nationality Netherland Dirk_Bouts nationality Netherland Netherland rdf:type Book

SELECT ?x ?y WHERE { ?x <http://dbpedia.org/ontology/nationality> ?y . ?y a <http://dbpedia.org/ontology/Book> . }

!nationality.Book # Person

Page 23: Statistical Schema Induction

!""#$%%&'(')*+,-."'&(/)'0-.))!1'-(213453453

<owl:Restriction> <rdfs:subClassOf rdf:resource="http://dbpedia.org/ontology/Wrestler"/> <owl:onProperty rdf:resource="http://dbpedia.org/ontology/trainer"/> <owl:someValuesFrom rdf:resource="http://dbpedia.org/ontology/ComicsCharacter"/> </owl:Restriction>

SELECT ?x ?y WHERE { ?x <http://dbpedia.org/ontology/trainer> ?y . ?y a <http://dbpedia.org/ontology/ComicsCharacter> . }

Jim_White_(wrestler) trainer Johnny_Thunder Johnny_Thunder rdf:type ComicsCharacter

!trainer.ComicsCharacter # Wrestler

Page 24: Statistical Schema Induction

!""#$%%&'(')*+,-."'&(/)'0-.))!1'-(213453453

{{Infobox mountain | name = San Benedicto | photo = San Benedicto Island - Landsat Image Cleaned.JPG | photo_caption = San Benedicto Island - Landsat Image N-12-15 2000 (1:25,000) | elevation = c.{{convert|332|m|ft|0}} | ... | last_eruption = [[1953]]<ref name="ReferenceA"/> | first_ascent = | easiest_route = }}

{{Infobox artist | bgcolour = #6495ED | name = Dirk Bouts | ... | birthplace = [[Haarlem]] | deathdate = ca. 1475 | deathplace = | nationality = [[Netherland]] | field = [[Painting]] | ... }}

{{ infobox wrestler | name = Jim White | ... | trainer = [[Johnny Thunder]] | debut = 1959<ref name=slam/> | retired = 1985<ref name=slam/> }}

Page 25: Statistical Schema Induction

!""#$%%&'(')*+,-."'&(/)'0-.))!1'-(213453453

62 479,132

64 133,149

29 1,138,601

20 1,757

41 14,322

Page 26: Statistical Schema Induction

!""#$%%&'(')*+,-."'&(/)'0-.))!1'-(213453453

Page 27: Statistical Schema Induction

!""#$%%&'(')*+,-."'&(/)'0-.))!1'-(213453453

support threshold 0.0 confidence threshold 1.0

FerryDockAccessArea # FerryStopPoint % O"StreetStopPoint % StopPointBusOrCoachStation # StopArea

City # NptgLocalityPrinciplePoint # StopPoint

HailAndRideBusStop #StopPoint % OnStreetStopPoint% BusStopPoint % BusCoachTramOnStreetPoint

Page 28: Statistical Schema Induction

!""#$%%&'(')*+,-."'&(/)'0-.))!1'-(213453453

�•

�• rdf:type

�•�•

�–�– rdf:type

�–

�– olivierAward # tonyAward

Page 29: Statistical Schema Induction

!""#$%%&'(')*+,-."'&(/)'0-.))!1'-(213453453

�•�–�–

�•�–�–

�–

* http://live.dbpedia.org

Page 30: Statistical Schema Induction

!""#$%%&'(')*+,-."'&(/)'0-.))!1'-(213453453

http://code.google.com/p/gold-miner/

!r." # C

r1 $ r2 # r3

C1 % C2 # C3

C1 # !r.C2

a2 & C1, C2, C5

a1 & C1, C4

a3 & C2, C3

Page 31: Statistical Schema Induction

!""#$%%&'(')*+,-."'&(/)'0-.))!1'-(213453453

<owl:Class rdf:about="http://reference.data.gov.uk/def/central-government/MinisterialDepartment"> <rdfs:subClassOf rdf:resource="http://reference.data.gov.uk/def/central-government/Department"/> <rdfs:subClassOf rdf:resource="http://reference.data.gov.uk/def/public-body/Department"/> </owl:Class>

<owl:Class rdf:about="http://reference.data.gov.uk/id/department/defra/grade/"> <rdfs:subClassOf rdf:resource="http://reference.data.gov.uk/def/central-government/CivilServicePost"/> </owl:Class>

<owl:Class rdf:about="http://reference.data.gov.uk/id/public-body/national-gallery/grade/1"> <rdfs:subClassOf rdf:resource="http://reference.data.gov.uk/def/central-government/CivilServicePost"/> </owl:Class>


Recommended