Title: Predictive subdata selection for large-scale deterministic computer models
Authors: Ming-Chung Chang - Academia Sinica (Taiwan) [presenting]
Abstract: Computer models are implementations of complex mathematical models using computer codes. Tremendous amounts of data generated from computer models are becoming ubiquitous owing to advanced technology. Such data richness, however, may yield an inability to conduct statistical analysis in terms of time cost. Recently, increased attention has focused on solving this data reduction problem. We will introduce a new subdata selection method for large-scale deterministic computer models. The proposed method takes advantage of the information of the output values and adaptively updates the current subdata with affordable computational cost. Simulated examples and real data analyses are provided.