A0191
Title: Group-orthogonal subsampling for big data linear mixed models
Authors: Fasheng Sun - Northeast Normal University (China) [presenting]
Abstract: The linear mixed model is a popular and common modelling method in statistical analysis. It is computationally challenging to obtain parameter estimates for big data in the linear mixed model. The current subsampling methods are aimed at the situation where the data is independent without considering the correlation within the data. An optimal subsampling method for a linear mixed model based on maximizing the determinant of the variance-covariance matrix of the subsampling estimator is proposed. The proposed subsampling procedure is also optimal under the $A$-optimality criterion, which minimizes the trace of the variance-covariance matrix of the subsampling estimator. Furthermore, the asymptotic property of the subsampling estimator is established. Numerical examples based on both simulated and real data are provided to illustrate the proposed subsampling method.