A novel ensemble learning approach to unsupervised record linkage
- Submitting institution
-
University of the West of England, Bristol
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 877940
- Type
- D - Journal article
- DOI
-
10.1016/j.is.2017.06.006
- Title of journal
- Information Systems
- Article number
- -
- First page
- 40
- Volume
- 71
- Issue
- -
- ISSN
- 0306-4379
- Open access status
- Compliant
- Month of publication
- June
- Year of publication
- 2017
- URL
-
https://doi.org/10.1016/j.is.2017.06.006
- Supplementary information
-
-
- Request cross-referral to
- -
- Output has been delayed by COVID-19
- No
- COVID-19 affected output statement
- -
- Forensic science
- No
- Criminology
- No
- Interdisciplinary
- No
- Number of additional authors
-
3
- Research group(s)
-
-
- Citation count
- 8
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- This is the first paper to propose a novel approach to unsupervised record linkage that combines a selected ensemble of automatically self-learned matchers for record linkage. This approach removes the requirement of large labelled training datasets in supervised record linkage and benefits from the diversity of matchers in the ensemble to improve the accuracy of individual self-learned matchers. It is one of the main outcomes of an ESRC funded Administrative Data Research Centre - Northern Ireland (ADRC-NI). This kind of unsupervised machine learning approach to record linkage can play an important role in administrative and public health data science applications.
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -