Entity Resolution for Longitudinal Studies in Education using OYSTER

摘要

This paper describes the application of Oyster, an open source, general purpose entity resolution (ER) system, to the problem of conducting multi-year longitudinal studies of student achievement. Although originally designed to support ER instruction and research, this paper demonstrates that OYSTER can be used in practical applications with processing requirements on the order of one million records, a range that includes many existing small-scale information systems. The paper also discusses an enhancement to the basic R-Swoosh algorithm implemented in OYSTER that allows higher performance in processing student files with high duplication rates.