Advances in record linkage: Linking historical data, string similarities and ethnicities, and consistently evaluating linkage
Export to calendarAbstract
In this presentation, Sumayya will discuss her research in record linkage, covering three interconnected areas: the linkage of historical population data, how ethnic name characteristics influence string similarities, and a novel approach to consistently evaluate record linkage performance.
Her work on historic linkage is associated with the Scottish Historic Population Platform (SHiPP) project at the University of Edinburgh. She will discuss how vital life events – such as births, deaths, and marriages – can be linked to census records using record linkage techniques to support the study of population dynamics and demographic change over time. The talk will highlight the methodologies used to link historical datasets, key findings from this work, and future research directions.
Sumayya will also present recent findings on the impact of different string comparison functions when linking populations with ethnically diverse names. Based on experiments conducted on a large modern population database, she will discuss how these choices can introduce bias and influence the linkage quality for different ethnic groups within the population.
Finally, she will discuss a novel approach for evaluating record linkage outcomes. This will highlight limitations and potential biases in existing evaluation methodologies and describe an alternative approach designed to provide more consistent and reliable assessments of linkage quality.
Register to attend online.
Bio
Sumayya Ziyad is a PhD candidate in the School of Computing at the Australian National University (ANU), where her research focuses on developing efficient privacy-preserving record linkage (PPRL) techniques with enhanced privacy guarantees. She holds a Bachelor’s degree with Honours from the University of Colombo, Sri Lanka, and brings several years of industry experience as a Software Engineer.
In addition to her research, Sumayya contributes to teaching at ANU as a tutor for the Data Wrangling course, supporting both undergraduate and postgraduate students. She has also been affiliated with the SHiPP project, contributing to research on historical population data and record linkage.
Event Details
Event Speakers
