项目反应理论中参数的双重两步迭代估计

Empirical Logistic Regression and Its Use in Estimating Parameters under Item Response Theory

将经验Logistic回归用于项目反应理论中两参数Logistic模型的参数估计,得到第j个项目的项目 ⌒-1 ⌒-1Xj)-1X′jYj,由于Xj含有未知的讨厌参数θ,我们结参数向量βj=(aj,λj)的两步估计β⌒⌒j=(X′jjj合上式的结果对θ进行再估计.从而形成一种新的估计方法—双重两步迭代估计.蒙特卡洛模拟结果显示,这种新的估计方法提高了估计对真值的恢复能力.

Empirical Logistic regression and its use in estimating parameters of two parameters Logistic model under Item Response Theory are introduced in this paper.We obtain the two-stage estimation β⌒⌒j=(X′j⌒-1jXj)-1XX′j⌒-1jYj of the item parameter vector βj of j the item.The above algorithm forms a double-two-stage iteration.The results of Monte Carlo stimulation show that this double-two-stage iteration algorithm is more effective than empirical Logistic regression after item and contributes to raise the capability of parameters recovery.