Learning Name Variants from Inexact High-Confidence Matches

G. Bloothooft; M.P. Schraagen

doi:10.1007/978-3-319-19884-2_4

Learning Name Variants from Inexact High-Confidence Matches

Research output: Chapter in Book/Report/Conference proceeding › Chapter › Academic › peer-review

Abstract

Name variants which differ more than a few characters can seriously hamper record linkage. A method is described by which variants of first names and surnames can be learned automatically from records that contain more information than needed for a true link decision. Post-processing and limited manual intervention (active learning) is unavoidable, however, to differentiate errors in the original and the digitised data from variants. The method is demonstrated on the basis of an analysis of 14.8 million records from the Dutch vital registration.

Original language	English
Title of host publication	Population Reconstruction
Editors	Gerrit Bloothooft, Peter Christen, Kees Mandemakers, Marijn Schraagen
Place of Publication	Cham
Publisher	Springer
Chapter	4
Pages	61-83
Number of pages	23
ISBN (Electronic)	978-3-319-19884-2
ISBN (Print)	978-3-319-19883-5
DOIs	https://doi.org/10.1007/978-3-319-19884-2_4
Publication status	Published - 4 Aug 2015

Keywords

record linkage
historical
names

Access to Document

10.1007/978-3-319-19884-2_4

chp%3A10.1007%2F978-3-319-19884-2_4Final published version, 731 KBLicence: Taverne

Cite this

@inbook{8ca44b4246ba4e4b9c8ca3502c1f5363,

title = "Learning Name Variants from Inexact High-Confidence Matches",

abstract = "Name variants which differ more than a few characters can seriously hamper record linkage. A method is described by which variants of first names and surnames can be learned automatically from records that contain more information than needed for a true link decision. Post-processing and limited manual intervention (active learning) is unavoidable, however, to differentiate errors in the original and the digitised data from variants. The method is demonstrated on the basis of an analysis of 14.8 million records from the Dutch vital registration.",

keywords = "record linkage, historical, names",

author = "G. Bloothooft and M.P. Schraagen",

year = "2015",

month = aug,

day = "4",

doi = "10.1007/978-3-319-19884-2_4",

language = "English",

isbn = "978-3-319-19883-5",

pages = "61--83",

editor = "Gerrit Bloothooft and Peter Christen and Mandemakers, {Kees } and Marijn Schraagen",

booktitle = "Population Reconstruction",

publisher = "Springer",

address = "Germany",

}

TY - CHAP

T1 - Learning Name Variants from Inexact High-Confidence Matches

AU - Bloothooft, G.

AU - Schraagen, M.P.

PY - 2015/8/4

Y1 - 2015/8/4

N2 - Name variants which differ more than a few characters can seriously hamper record linkage. A method is described by which variants of first names and surnames can be learned automatically from records that contain more information than needed for a true link decision. Post-processing and limited manual intervention (active learning) is unavoidable, however, to differentiate errors in the original and the digitised data from variants. The method is demonstrated on the basis of an analysis of 14.8 million records from the Dutch vital registration.

AB - Name variants which differ more than a few characters can seriously hamper record linkage. A method is described by which variants of first names and surnames can be learned automatically from records that contain more information than needed for a true link decision. Post-processing and limited manual intervention (active learning) is unavoidable, however, to differentiate errors in the original and the digitised data from variants. The method is demonstrated on the basis of an analysis of 14.8 million records from the Dutch vital registration.

KW - record linkage

KW - historical

KW - names

U2 - 10.1007/978-3-319-19884-2_4

DO - 10.1007/978-3-319-19884-2_4

M3 - Chapter

SN - 978-3-319-19883-5

SP - 61

EP - 83

BT - Population Reconstruction

A2 - Bloothooft, Gerrit

A2 - Christen, Peter

A2 - Mandemakers, Kees

A2 - Schraagen, Marijn

PB - Springer

CY - Cham

ER -

Learning Name Variants from Inexact High-Confidence Matches

Abstract

Keywords

Access to Document

Fingerprint

Cite this