A0569
Title: A study on estimation in multivariate allometric regression
Authors: Koji Tsukuda - Kyushu University (Japan) [presenting]
Shun Matsuura - Keio University (Japan)
Abstract: The multivariate allometric regression model is an extension of the allometric extension model to multivariate regression, which has been proposed previously. In the multivariate allometric regression model, it is assumed that the direction of the difference between the mean vectors of response variables for different values of explanatory variables always coincides with the first principal eigenvector of the covariance matrix of errors. Some estimators of the first principal eigenvector based on preliminary tests are proposed. The proposed estimators are also compared with conventional estimators.