A0942
Title: Local linear estimation for covariate-dependent coefficients model in disease mapping
Authors: Pei-Sheng Lin - National Health Research Institutes (Taiwan)
Jun Zhu - University of Wisconsin (United States)
Yexuan Jiang - University of North Carolina at Chapel Hill (United States)
Feng-Chang Lin - University of North Carolina at Chapel Hill (United States) [presenting]
Abstract: Spatial regression effects may depend on covariates in the disease mapping modeling. Taking infectious disease, for example, the association between incidence and risk factors may vary by climatic factors such as season and temperature. The aim is to build a model that can simultaneously study spatial and varying effects of covariates that are deemed to modulate the spatial association. A local linear estimation method is employed for the covariate-dependent coefficients in a spatial model dealing with excess zero counts. The local linear estimator is designed to smooth time-dependent coefficient estimation effectively. Comprehensive simulation studies were used to demonstrate the performance of our local linear estimators under various scenarios. Dengue incidences in the villages using weekly reported dengue cases in Kaohsiung City from January 2014 to December 2015 are used to offer insights into the proposed method for the practical use of real-world applications.
