A longitudinal pattern based approach to predict Alzheimer's disease via Lead Investigator: Ying Lin Institution : University of Washington E-Mail : liny90@uw.edu Proposal ID : 907 Proposal Description: Identifying the individuals who will progress to AD in an early stage is critical for initializing the preventive care and treatment of AD. Predicting the individuals' disease progression from the interactive, longitudinal, and multivariate measurements, e.g. clinical measurements, neuropsychological tests, and health history, is challenging. We aim to develop a rule-based method to 1) discover a set of risk predictive patterns to capture both progression trajectory in individual measurement and interactions between multiple measurements 2) fusing these patterns to assess the individual's risk of disease onset. The overall goal is to use the multivariate longitudinal measurements for accurately predicting AD.