Data Integration Framework: A Children and Parents Cohort Case Study

This paper presents a proposal for a data integration framework. The purpose of the framework is to locate automatically records of participants from the ALSPAC database (Avon Longitudinal Study of Parents and Children) within its counterpart GPRD database (General Practice Research Database). The ALSPAC database is a collection of data from children and parents from before birth to late puberty. This collection contains several variables of interest for clinical researchers but we concentrate in asthma as a golden standard for evaluation of asthma has been made by a clinical researcher. The main component of the framework is a module called Mapper which locates similar records and performs record linkage. The mapper contains a library of similarity measures such Jaccard, Jaro-Winkler, Monge-Elkan, MatchScore, Levenstein and TFIDF similarity. Finally, the author evaluates the approach on quality of the mappings.