ASYMMETRIC ASYMPTOTIC CONFIDENCE REGION FOR A MULTIVARIATE LINEAR STRUCTURAL RELATIONSHIP

抄録

For the problem of finding a confidence region for a multivariate linear structural relation, Okamoto & Isogawa [7] proposed a method to construct an asymptotic confidence region as the sample size tends to infinity by treating the true variables symmetrically. This article presents a method to construct another asymptotic confidence region by treating the variables asymmetrically.