Title: Asymptotic efficiency of Cp-type criterion in high-dimensional multivariate linear regression models
Authors: Shinpei Imori - Hiroshima University (Japan) [presenting]
Abstract: Variable selection is one of the crucial problems in statistical fields. Consistency and asymptotic efficiency are known as typical asymptotic properties of selection methods. In univariate linear regression models, Cp criterion is a feasible method from the perspective of prediction because it has the asymptotic efficiency, whereas it does not have the consistency. Recently, the consistency property of Cp-type criterion is shown in high-dimensional multivariate linear regression models, where the dimension of a response matrix is large. We study the asymptotic efficiency of Cp-type criterion under the high-dimensional framework.