A0627
Title: Variable selection in high-dimensional multivariate linear regression models with group structure
Authors: Shinpei Imori - Hiroshima University (Japan) [presenting]
Abstract: A variable selection problem is studied in high-dimensional multivariate linear regression models. In multivariate linear regression models, it is often assumed to have common explanatory variables for each response variable. However, the condition that the same explanatory variables are used for each response variable cannot express group structure among response variables. We mitigate this assumption and derive a sufficient condition so that Cp-type criteria have asymptotic efficiency when the number of response variables and explanatory variables is large.
