项目无响应情况下MIRT模型潜变量选择的EMS算法  

EMS algorithm for latent variable selection in MIRT model with item nonresponse

作　　者：刘芙蓉 徐平峰 LIU Furong;XU Pingfeng(School of Mathematics&Statistics,Changchun University of Technology,Changchun 130012,China)

机构地区：[1]长春工业大学数学与统计学院,吉林长春130012

出　　处：《长春工业大学学报》2024年第6期537-543,共7页Journal of Changchun University of Technology

基　　金：吉林省自然科学基金项目(20210101152JC)。

摘　　要：考虑在教育评估和心理测试中,项目无响应情况下潜变量选择的问题。针对缺失数据过程建立一个项目反应理论(IRT)模型,应用期望模型选择(EMS)算法对其进行潜变量选择。模拟显示,此方法参数估计和潜变量选择结果均较为准确。并用其分析2018国际学生评估计划(PISA)针对学生进行的问卷调查数据。Consider the problem of latent variable selection in the case of item unresponsiveness in educational assessments and psychological tests.An item response theory(IRT)model is established for the missing data process,and the expected model selection(EMS)algorithm is used to select the latent variables.The simulation results show that the results of the parameter estimation and latent variable selection of this method are accurate.It was used to analyze data from the 2018 Programme for International Student Assessment(PISA)survey of students.

关 键 词：MIRT模型 缺失响应 EMS算法 

分 类 号：O21[理学—概率论与数理统计]