ChartEx: Tools for the Analysis of Medieval ChartersAdam Kosto, Bob Scott (Columbia University) Bob...
Transcript of ChartEx: Tools for the Analysis of Medieval ChartersAdam Kosto, Bob Scott (Columbia University) Bob...
ChartEx: Tools for the Analysis of Medieval
Charters
www.chartex.org
Adam Kosto Columbia University
Universität Zürich 10 November 2016
Project Members / Institutions Sarah Rees Jones, Helen Petrie, Chris Power,
Stefania Perring (University of York) Roger Evans, Lynne Cahill (University of Brighton) Adam Kosto, Bob Scott (Columbia University) Bob Stacey, Jon Crump (University of
Washington) Arno Knobbe, Marvin Meeng (Leiden
University) Michael Gervers, Robin Sutherland-Harris
(University of Toronto)
Sale of Property D, bounded by A, B, and C
to South
Sale of Property B, bounded by Property C
to East
Sale of Property A, bounded by Property B
to East
Document 1 Document 2 Document 3
A B C
D
N
ChartEx Architecture
charters
analysed individual charters
NLP Data mining
analysed integrated charters
Workbench
ChartEx Developmental Data Charters of the Vicars Choral of York Minster
Latin charters and English abstracts.
The National Archives (UK) Ward 2
English abstracts
The Borthwick Institute, Yarburgh Muniments
English abstracts
DEEDS (University of Toronto)
Latin charters of English provenance
Cluny (CBMA)
Latin charters of French provenance
BRAT Rapid Annotation Tool
Pontus Stenetorp, Sampo Pyysalo, Goran Topić, Tomoko Ohta, Sophia Ananiadou and Jun'ichi Tsujii (2012). brat: a Web-‐based Tool for NLP-‐Assisted Text AnnotaPon. In Proceedings of the Demonstra6ons Session at EACL 2012. (hRp://brat.nlplab.org/)
Ronald Denaux , Catherine Dolbear , Glen Hart , Vania Dimitrova , Anthony G. Cohn, “SupporPng domain experts to construct conceptual ontologies: A holisPc approach,” Web Seman6cs: Science, Services and Agents on the World Wide Web Volume 9.2 (2011): 113 -‐ 127
Kanga Ontology Development
Simplified ChartEx Markup Schema ENTITIES
Actors (e.g., John, Thomas)
Locations (e.g., Blackacre, London) Events (e.g., sale, grant) Occupation (e.g., Smith) Date (e.g., 1 May 1200)
Simplified ChartEx Markup Schema ENTITIES
Actors (e.g., John, Thomas)
Locations (e.g., Blackacre, London) Events (e.g., sale, grant) Occupation (e.g., Smith) Date (e.g., 1 May 1200)
ROLES
is grantor of is father or
is located W of occurs on
is the same as
Simplified ChartEx Markup Schema ENTITIES
Actors (e.g., John, Thomas)
Locations (e.g., Blackacre, London) Events (e.g., sale, grant) Occupation (e.g., Smith) Date (e.g., 1 May 1200)
ROLES
is grantor of is father or
is located W of occurs on
is the same as
RELATIONSHIPS BETWEEN ENTITIES (TRIPLES) ENTITY --- ROLE --- ENTITY
John is the father of Thomas grant occurs on 1 May 1200 Thomas is the grantor of Blackacre Blackacre is located W of London
Thomas Smith of York, son of Hugh
Thomas of York, smith, son of Hugh
Thomas fitz Hugh, of York, smith
Marking Up Persons
Thomas Smith of York, son of Hugh
Thomas of York, smith, son of Hugh
Thomas fitz Hugh, of York, smith
Marking Up Persons
Thomas Smith of York, son of Hugh is son of
Thomas of York, smith, son of Hugh
Thomas fitz Hugh, of York, smith
Marking Up Persons is from
Marking Up Places
Bishop John grants income from land in the hundred of Wells to Glastonbury Abbey, namely from the woods called the Grava bounded to the north by a stream and to the south by Robert son of Adam’s farm, and from….
Marking Up Places
Bishop John grants income from land in the hundred of Wells to Glastonbury Abbey, namely from the woods called the Grava bounded to the north by a stream and to the south by Robert son of Adam’s farm, and from….
Marking Up Places
Bishop John grants income from land in the hundred of Wells to Glastonbury Abbey, namely from the woods called the Grava bounded to the north by a stream and to the south by Robert son of Adam’s farm, and from….
Bishop John grants income from land in the hundred of Wells to Glastonbury Abbey, namely from the woods called the Grava bounded to the north by a stream and to the south by Robert son of Adam’s farm, and from….
Charter Annotation
Inter-Charter Probability • Statistics
p(Thomas) = 0.12 (common name) p(Josce) = 0.0015 (uncommon name) p(Goldsmith) = 0.04 (common profession)
• Dating
vc-408 1252-1253 vc-409 1253-1261
• Final confidence
conf (Thomas 408, Thomas 409) = 0.9993 i.e., there is a 99% chance that the Thomas in doc. 408
and the Thomas in doc. 409 are the same person