Special Colloquium
Zhengyuan Zhu
Iowa State University
Some Theory and Practice of Statistical Data Integration
Abstract: Combining information from multiple sources to make valid statistical inference is of great practical importance in many fields. In this talk I will focus on applications in agricultural and environmental statistics and share our experience integrating remote sensing data, administrative data, and survey data to provide better estimates at smaller spatial scales. New spatial-temporal functional imputation method for satellite data gap-filling, machine learning methods for satellite data based land-cover classification, hierarchical area-level model and spatial sub-group analysis, and functional covariate balancing methods will be introduced, which provides a theoretically justifiable tool box for statistical data integration. Case studies in agricultural and natural resource surveys will be presented to illustrate the success of the methods and the remaining challenges.
Tea will be served at 4:00 pm in SEO 300.
Wednesday February 6, 2019 at 3:00 PM in 636 SEO