Publications by authors named "Clicia Pinto"
On the Accuracy and Scalability of Probabilistic Data Linkage Over the Brazilian 114 Million Cohort.
IEEE J Biomed Health Inform
March 2018
Article Synopsis
- Data linkage involves connecting records referring to the same entity across different data sources, essential for research in fields like public health.
- When unique identifiers are lacking, probabilistic methods are used to assess similarities between records, which requires careful selection of attributes and metrics.
- The paper introduces AtyImo, a hybrid probabilistic linkage tool that shows high accuracy (93%-97% true matches) and can efficiently process large datasets, linking 114 million individuals in Brazil in under nine days.