A0662
Title: A new perspective on unsupervised learning in longitudinal studies
Authors: Hyunkeun Cho - University of Iowa (United States) [presenting]
Daren Kuwaye - University of Iowa (United States)
David-Erick Lafontant - University of Iowa (United States)
Abstract: The most common use of unsupervised learning is to cluster data into homogeneous groups. As units are measured multiple times, clustering longitudinal data has become popular. Over the last two decades, various clustering methods have been developed to cluster units based on the similarity of longitudinal profiles. new paradigm in longitudinal clustering is introduced, and its potential applications and implications for other clustering methods are discussed. This novel clustering method provides a unique perspective on longitudinal clustering by considering data heterogeneity over time.