Family History Research Family History Research : on the Semantic Web : on the Semantic Web Building a Semantic Prototype for Danish Building a Semantic Prototype for Danish Genealogical Research Genealogical Research By By Charla Woodbury and David W. Embley Charla Woodbury and David W. Embley BYU Computer Science Department BYU Computer Science Department charlajw@cs.byu.edu .edu embley@cs.byu.edu embley@cs.byu.edu charlajw@cs.byu Family History Technology Institute Family History Technology Institute March 24, 2005 March 24, 2005 Supported in part by NSF Supported in part by NSF
Semantic Web Semantic Web Machine “Understandable” Web Machine “Understandable” Web MEANING KNOWLEDGE INFORMATION DATA 2
Need for Semantic Web Need for Semantic Web “The Semantic Web: … content that is meaningful to computers [and that] will unleash a revolution of new possibilities … Properly designed, the Semantic Web can assist the evolution of human knowledge …” (Tim Berners-Lee, …, Weaving the Web) 3
Semantic Web Semantic Web ‘DATE DATE’ ’ ‘ Calendar date To date an artifact A fruit A romantic experience To go on a romantic experience with someone 4
Also a SURNAME SURNAME – – Also a Mr. C. J. Date** ** Mr. C. J. Date The semantic web will make it possible for machines to know the difference! ** Edgar F. Codd and C. J. Date are famous in the area of databases for defining levels of normal forms 5
Real Problem Real Problem A person decides to do family history research for the first time on their Danish family lines. • Where do they go? • What records do they look for? • How do they handle records in Danish? • How can they tell when the records they have match their search family? 6
Semantic Web Semantic Web Ideal for Family History Ideal for Family History SOLUTION – PROTOTYPE The heart of a one-stop web site for naïve researchers So many records have been extracted into digitized forms and are often available on the Web Limited geographically – parish and probate records from Nim District, Skanderborg, Denmark • 100% probates 100% probates • 100% marriages 100% marriages 7
Semantic Web Prototype Ontology – semantic model (BYU Ontos) Annotated web pages (Web Ontology Language OWL proposed W3C Feb 2004) Solutions for special genealogical problems 8
Ontology Model Ontology Model 9
Person Matching Person Matching in genealogical research in genealogical research NAMES DATES PLACES RELATIONS 10
Ontology Entities Ontology Entities FIND and MARK UP relevant web pages by: • NAME NAME <NAME> <NAME> • DATE DATE <DATE> <DATE> • PLACE PLACE <PLACE> <PLACE> • RELATIONSHIP RELATIONSHIP <RELATION> <RELATION> • OCCUPATION OCCUPATION <OCCUPATION> <OCCUPATION> • RECORD_TYPE RECORD_TYPE <RTYPE> <RTYPE> • SOURCE SOURCE <SOURCE> <SOURCE> 11
Partial Danish GIVEN NAME GIVEN NAME Partial Danish LEXICON LEXICON MALE FEMALE • And. And. • Ane Ane • Anders Anders • Anna Anna • Andreas Andreas • Anne Anne • Christen Christen • Birthe Birthe • Christian Christian • Birte Birte • Eric Eric • Bodil Bodil • Erik • Caroline Erik Caroline • Gregers Gregers • Dorte Dorte • Hans Hans • Dorthe Dorthe • Ib • Elene Ib Elene • Jacob Jacob • Ellen Ellen • Jens Jens • Elisabeth Elisabeth • Jep Jep • Elsbeth Elsbeth 12
Partial DATE DATE Lexicon Lexicon Partial (actual lexicon is a single list in alphabetic order) (actual lexicon is a single list in alphabetic order) FEAST DATES (partial) MONTHS January –Jan –Januar -11br Easter – Paaske –Påske –Paasche –Påsche Februrary –Feb –Februar -12br March –Mar –Marts Pentecost – Pent –Pinse -Pin April – Apr –Apl Trinity –Tr –Trin –Trinitatis May –Mai June –Jun –Juni July –Jul –Juli -5br DAYS OF WEEK August –Aug –Augst -6br September –Sep –Sept -7br Sunday –Dominico –Dom. –Septembre Monday –Mondag –Mond. October –Oct -8br –Octobre Tuesday –Tirsdag –Tirsd. November –Nov -9br –Novembre Wednesday -Onsdag –Onsd. December –Dec -10br -Decembre Thursday –Tørsdag –Tørsd. Friday –Fredag –Fred. TIME Saturday –Lørsdag –Lørs. Year –yr –aar –år Month –mo –maaned –måned –m. Week –uge –ug. Day –dag –dg. Hour – h. –hr. 13
Original Record Original Record FHL Film#052,236 Tvilum Parish FHL Film#052,236 Tvilum Parish 14
Web Page Web Page • SOURCE URL - SOURCE URL -Tvilum Sogne Kirkebog Tvilum Sogne Kirkebog • [PAGE HEADER] [PAGE HEADER] Fødde 1751 3 Fødde 1751 3 • [BODY] [BODY] Truust Dom. 23 p: Trinit: laest over Niels Truust Dom. 23 p: Trinit: laest over Niels Baches SØREN fadd. Johannes Michelsens og Niels Baches SØREN fadd. Johannes Michelsens og Niels Mollers hustruer af Søebyevad, Peder Rasmussen af Mollers hustruer af Søebyevad, Peder Rasmussen af Søebyevad, Jens Bachis søn Peder og Niels Thylkes s. Søebyevad, Jens Bachis søn Peder og Niels Thylkes s. Peder af Truust Peder af Truust 15
Ontology Entities Ontology Entities FIND and MARK UP relevant web pages by: • NAME NAME <NAME> <NAME> • DATE DATE <DATE> <DATE> • PLACE PLACE <PLACE> <PLACE> • RELATIONSHIP RELATIONSHIP <RELATION> <RELATION> • OCCUPATION OCCUPATION <OCCUPATION> <OCCUPATION> • RECORD_TYPE RECORD_TYPE <RTYPE> <RTYPE> • SOURCE SOURCE<SOURCE> <SOURCE> Colors only represent OWL annotation mark-ups Colors only represent OWL annotation mark-ups automatically placed in the web page using the ontology automatically placed in the web page using the ontology 16
Annotated Web Page Annotated Web Page • SOURCE - SOURCE -Tvilum Parish Register Tvilum Parish Register • [PAGE HEADER] [PAGE HEADER] Fødde 1751 1751 3 3 Fødde • [BODY] [BODY] Truust Truust Dom. 23 p: Trinit: Dom. 23 p: Trinit: laest over laest over Niels Niels Baches SØREN SØREN fadd. fadd. Johannes Michelsens Johannes Michelsens og og Niels Niels Baches Mollers hustruer hustruer af af Søebyevad Søebyevad, , Peder Rasmussen Peder Rasmussen af af Mollers Søebyevad, , Jens Bachis Jens Bachis søn søn Peder Peder og og Niels Thylkes Niels Thylkes s. s. Søebyevad Peder af af Truust Truust Peder 17
Results Listing Results Listing TARGET – Jens Pedersen Bach Truust, Tvilum Parish, Gjern District, Skanderborg Date Range - born 1693 to died 1778 Name Date Place Relation Occupation Record Source Type (URL) Jens Bachis Dom. 23 p: Truust fadd: fadd: Fødde Fødde Tvilum Trinit: Parish 1751 Register (14 Nov 1751) SOURCE -Tvilum Parish Register Tvilum Parish Register SOURCE - [PAGE HEADER] Fødde Fødde 1751 1751 3 3 [PAGE HEADER] [BODY] Truust Truust Dom. 23 p: Trinit: Dom. 23 p: Trinit: laest over laest over Niels Baches Niels Baches SØREN SØREN [BODY] fadd. Johannes Michelsens Johannes Michelsens og og Niels Niels Mollers Mollers hustruer hustruer af af Søebyevad Søebyevad, , fadd. Jens Bachis , Jens Bachis Peder Rasmussen af af Søebyevad Søebyevad, søn søn Peder Peder og og Peder Rasmussen Niels Thylkes s. s. Peder Peder af af Truust Truust Niels Thylkes 18
Conversion Functions Conversion Functions inside the ontology inside the ontology • Compute birthdate from age at death Compute birthdate from age at death Death – 22 Mar 1743 Death – 22 Mar 1743 Age - 23 yr 2 m Age - 23 yr 2 m -> BIRTH -> BIRTH Jan 1720 Jan 1720 • Compute dates from feast dates Sunday 23 rd after Trinity 1751 -> 14 Nov 1751 19
Solutions for Special Problems Solutions for Special Problems RULES FOR • Matching different name forms Matching different name forms • Matching place names to appropriate records Matching place names to appropriate records 20
RULE - Match different name forms - Match different name forms RULE as ONE PERSON as ONE PERSON • JENS PEDERSEN JENS PEDERSEN • JENS PEDERSEN BACH JENS PEDERSEN BACH • JENS BACH JENS BACH • JENS BACHIS JENS BACHIS 21
PLACES - County Map of DENMARK 22
Parish and District Map of Parish and District Map of SKANDERBORG SKANDERBORG 23
Road Map Road Map www.expedia.com www.expedia.com 24
Matching Places to Records Matching Places to Records Farm Parish District County Record Links name Molger Tamdrup Nim Skanderborg PARISH Tamdrup 1684-1912 PROBATE Nim Herred Provisti Rask Skanderborg Rytterdistrikt Tamdrup Nim Skanderborg List of URL’s Includes Molger URL’s Adds Parish specific records Nim Skanderborg List of URL’s Includes Tamdrup URL’s Adds District specific records Skanderborg List of URL’s Includes all district URL’s Adds County specific records 25
Recommend
More recommend
