关键词: 半相依模型, 协方差改进估计, 两步估计

Abstract:

For a seemingly unrelated regression with m linear regression models,the literature [7] propose a new method which based on iterative information step by step. Based on this method, this paper obtain a farther adjusted form which also has congruence. At the same time, a new estimator is proposed by using the idea in literature [7]. It is a feasible estimator having small sample property. The simulation result also indicates that our estimator is effective. The literature [10] gave a concise two-stage estimator by using the covariance adjusted approach of
literature [8]. This paper improves on it which shows the more power of small sample optimality by comparing with least square estimator.

Key words: Seemingly unrelated regression model, Covariance adjusted estimator, Two-stage estimator