Multiple instance learning for group record linkage

Zhichun Fu; Jun Zhou; Peter Christen; Mac Boot

doi:10.1007/978-3-642-30217-6_15

Multiple instance learning for group record linkage

Zhichun Fu^*, Jun Zhou, Peter Christen, Mac Boot

^*Corresponding author for this work

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

13 Citations (Scopus)

Abstract

Record linkage is the process of identifying records that refer to the same entities from different data sources. While most research efforts are concerned with linking individual records, new approaches have recently been proposed to link groups of records across databases. Group record linkage aims to determine if two groups of records in two databases refer to the same entity or not. One application where group record linkage is of high importance is the linking of census data that contain household information across time. In this paper we propose a novel method to group record linkage based on multiple instance learning. Our method treats group links as bags and individual record links as instances. We extend multiple instance learning from bag to instance classification to reconstruct bags from candidate instances. The classified bag and instance samples lead to a significant reduction in multiple group links, thereby improving the overall quality of linked data. We evaluate our method with both synthetic data and real historical census data.

Original language	English
Title of host publication	Advances in Knowledge Discovery and Data Mining - 16th Pacific-Asia Conference, PAKDD 2012, Proceedings
Pages	171-182
Number of pages	12
Edition	PART 1
DOIs	https://doi.org/10.1007/978-3-642-30217-6_15
Publication status	Published - 2012
Event	16th Pacific-Asia Conference on Advances in Knowledge Discovery and Data Mining, PAKDD 2012 - Kuala Lumpur, Malaysia Duration: 29 May 2012 → 1 Jun 2012

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Number	PART 1
Volume	7301 LNAI
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	16th Pacific-Asia Conference on Advances in Knowledge Discovery and Data Mining, PAKDD 2012
Country/Territory	Malaysia
City	Kuala Lumpur
Period	29/05/12 → 1/06/12

Access to Document

10.1007/978-3-642-30217-6_15

Cite this

Fu, Z., Zhou, J., Christen, P., & Boot, M. (2012). Multiple instance learning for group record linkage. In Advances in Knowledge Discovery and Data Mining - 16th Pacific-Asia Conference, PAKDD 2012, Proceedings (PART 1 ed., pp. 171-182). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 7301 LNAI, No. PART 1). https://doi.org/10.1007/978-3-642-30217-6_15

Fu, Zhichun ; Zhou, Jun ; Christen, Peter et al. / Multiple instance learning for group record linkage. Advances in Knowledge Discovery and Data Mining - 16th Pacific-Asia Conference, PAKDD 2012, Proceedings. PART 1. ed. 2012. pp. 171-182 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 1).

@inproceedings{1b59028006154cc7a2942598f0918524,

title = "Multiple instance learning for group record linkage",

abstract = "Record linkage is the process of identifying records that refer to the same entities from different data sources. While most research efforts are concerned with linking individual records, new approaches have recently been proposed to link groups of records across databases. Group record linkage aims to determine if two groups of records in two databases refer to the same entity or not. One application where group record linkage is of high importance is the linking of census data that contain household information across time. In this paper we propose a novel method to group record linkage based on multiple instance learning. Our method treats group links as bags and individual record links as instances. We extend multiple instance learning from bag to instance classification to reconstruct bags from candidate instances. The classified bag and instance samples lead to a significant reduction in multiple group links, thereby improving the overall quality of linked data. We evaluate our method with both synthetic data and real historical census data.",

keywords = "Multiple instance learning, entity resolution, historical census data, instance classification, record linkage",

author = "Zhichun Fu and Jun Zhou and Peter Christen and Mac Boot",

year = "2012",

doi = "10.1007/978-3-642-30217-6_15",

language = "English",

isbn = "9783642302169",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

number = "PART 1",

pages = "171--182",

booktitle = "Advances in Knowledge Discovery and Data Mining - 16th Pacific-Asia Conference, PAKDD 2012, Proceedings",

edition = "PART 1",

note = "16th Pacific-Asia Conference on Advances in Knowledge Discovery and Data Mining, PAKDD 2012 ; Conference date: 29-05-2012 Through 01-06-2012",

}

Fu, Z, Zhou, J, Christen, P & Boot, M 2012, Multiple instance learning for group record linkage. in Advances in Knowledge Discovery and Data Mining - 16th Pacific-Asia Conference, PAKDD 2012, Proceedings. PART 1 edn, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), no. PART 1, vol. 7301 LNAI, pp. 171-182, 16th Pacific-Asia Conference on Advances in Knowledge Discovery and Data Mining, PAKDD 2012, Kuala Lumpur, Malaysia, 29/05/12. https://doi.org/10.1007/978-3-642-30217-6_15

Multiple instance learning for group record linkage. / Fu, Zhichun; Zhou, Jun; Christen, Peter et al.
Advances in Knowledge Discovery and Data Mining - 16th Pacific-Asia Conference, PAKDD 2012, Proceedings. PART 1. ed. 2012. p. 171-182 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 7301 LNAI, No. PART 1).

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

TY - GEN

T1 - Multiple instance learning for group record linkage

AU - Fu, Zhichun

AU - Zhou, Jun

AU - Christen, Peter

AU - Boot, Mac

PY - 2012

Y1 - 2012

N2 - Record linkage is the process of identifying records that refer to the same entities from different data sources. While most research efforts are concerned with linking individual records, new approaches have recently been proposed to link groups of records across databases. Group record linkage aims to determine if two groups of records in two databases refer to the same entity or not. One application where group record linkage is of high importance is the linking of census data that contain household information across time. In this paper we propose a novel method to group record linkage based on multiple instance learning. Our method treats group links as bags and individual record links as instances. We extend multiple instance learning from bag to instance classification to reconstruct bags from candidate instances. The classified bag and instance samples lead to a significant reduction in multiple group links, thereby improving the overall quality of linked data. We evaluate our method with both synthetic data and real historical census data.

AB - Record linkage is the process of identifying records that refer to the same entities from different data sources. While most research efforts are concerned with linking individual records, new approaches have recently been proposed to link groups of records across databases. Group record linkage aims to determine if two groups of records in two databases refer to the same entity or not. One application where group record linkage is of high importance is the linking of census data that contain household information across time. In this paper we propose a novel method to group record linkage based on multiple instance learning. Our method treats group links as bags and individual record links as instances. We extend multiple instance learning from bag to instance classification to reconstruct bags from candidate instances. The classified bag and instance samples lead to a significant reduction in multiple group links, thereby improving the overall quality of linked data. We evaluate our method with both synthetic data and real historical census data.

KW - Multiple instance learning

KW - entity resolution

KW - historical census data

KW - instance classification

KW - record linkage

UR - http://www.scopus.com/inward/record.url?scp=84861452098&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-30217-6_15

DO - 10.1007/978-3-642-30217-6_15

M3 - Conference contribution

SN - 9783642302169

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 171

EP - 182

BT - Advances in Knowledge Discovery and Data Mining - 16th Pacific-Asia Conference, PAKDD 2012, Proceedings

T2 - 16th Pacific-Asia Conference on Advances in Knowledge Discovery and Data Mining, PAKDD 2012

Y2 - 29 May 2012 through 1 June 2012

ER -

Fu Z, Zhou J, Christen P, Boot M. Multiple instance learning for group record linkage. In Advances in Knowledge Discovery and Data Mining - 16th Pacific-Asia Conference, PAKDD 2012, Proceedings. PART 1 ed. 2012. p. 171-182. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 1). doi: 10.1007/978-3-642-30217-6_15

Multiple instance learning for group record linkage

Abstract

Publication series

Conference

Access to Document

Other files and links

Fingerprint

Cite this