|
Abstract:
|
Computer models are commonly used to represent a wide range of real systems, but they often involve some unknown parameters. Estimating the parameters by collecting physical data becomes essential in many scientific fields, ranging from engineering to biology. However, most of the existing methods are developed under the assumption that the physical data contains homoscedastic measurement errors. Motivated by an experiment of plant relative growth rates where replicates are available, we propose a new calibration method for inexact computer models with heteroscedastic measurement errors. Asymptotic properties of the parameter estimators are derived, and a goodness-of-fit test is developed to detect the presence of heteroscedasticity. Numerical examples and empirical studies demonstrate that the proposed method not only yields accurate parameter estimation, but it also provides accurate predictions for physical data in the presence of both heteroscedasticity and model misspecification.
|
