Predicting the Spatio-Temporal Evolution of Chronic Diseases in Population with Human Mobility Data

摘要: © 2018 International Joint Conferences on Artificial Intelligence. All right reserved. Chronic diseases like cancer and diabetes are major threats to human life. Understanding the distribution progression of chronic a population is important in assisting allocation medical resources as well design policies preemptive healthcare. Traditional methods obtain large scale indicators health, e.g., surveys statistical analysis, can be costly time-consuming often lead coarse spatio-temporal picture. In this paper, we leverage dataset describing mobility patterns citizens metropolitan area. By viewing local lifestyles predict evolution rate several at level city neighborhood. We apply combination collaborative topic modeling (CTM) Gaussian mixture method (GMM) tackle data sparsity challenge achieve robust predictions health conditions simultaneously. Our enables analysis prediction disease fine scales demonstrates potential incorporating datasets from mobile web sources improve monitoring. Evaluations using real-world check-in morbidity London show that proposed CTM+GMM model outperforms various baseline methods.