B1059
Title: A mixed effects model for cylindrical data with application to small area estimation
Authors: Tsubasa Ito - Hokkaido University (Japan)
Shogo Kato - Institute of Statistical Mathematics (Japan) [presenting]
Abstract: Cylindrical data are a set of bivariate observations that are paired between a linear variable and a circular variable. A mixed effects model for cylindrical data is proposed. For the proposal of the new model, a distribution for cylindrical data is presented, and then it is applied to define the mixed effects model with cylindrical responses and multivariate covariates. Some topics related to the proposed mixed effects model are discussed such as a measure of intra-cluster dependence between linear and circular variables, prediction of the random effects, and parameter estimation. The mixed effects model is applied to small area estimation, and the empirical best predictors of the small area mean and mean direction are derived. Finally, the proposed methods are demonstrated through simulation studies and a real dataset in biology.