在配適相關性資料時，關聯結構（Copula）模型是一種目前常用的方法。本文探討在廣義線性模型的假設下，當關聯結構模型假設錯誤時，迴歸參數之估計量的一致性問題。同時我們也將二元負二項模型的結果與關聯結構模型結果做對比。;Copula models are popular in modeling correlated data. This research investigates the performance of Copula models when model assumption fails. In the setting of generalized linear models we will show that regression parameter estimates are sensitive to model misspecification and provide an alternative approach to inference about regression parameters without knowing the true underlying distribution.
