A0490
Title: Dynamic design of experiments for function-on-function linear models
Authors: Caterina May - Kings College London (United Kingdom) [presenting]
Abstract: The optimal design of experiments in the context of functional data has been explored very little until now. Linear models are considered where both the response and one or more factors are continuous functions, for example, of the time. The goal is to find the optimal dynamic experimental conditions to estimate precisely the functional coefficients of the model. After establishing a suitable estimator and obtaining its variance-covariance matrix, the definition of optimal design criteria is extended to this functional context. A-optimal and D-optimal functional designs are then computed in practice, using the choice of suitable bases of functions to represent the data. The efficiency of estimators obtained under different choices is then compared.
