Subpopulation Discovery and Validation in Epidemiological Data

Motivated by identifying subpopulations that share common characteristics (eg alcohol consumption) to explain risk factors of diseases in cohort study data, we used subspace clustering to discover such subpopulations. In this paper, we describe our interactive coordinated multiple view system Visual Analytics framework S-ADVIsED for SubpopulAtion Discovery and Validation In Epidemiological Data. S-ADVIsED enables epidemiologists to explore and validate findings derived from subspace clustering. We investigated the replication of a selected subpopulation in an independent population.