双题匹配的MPI方法及其在平行试卷生成中的应用  

The Maximum Priority Index Method of Two Items Matching and Its Application on Constructing Parallel Tests

作　　者：汪文义[1] 熊建 宋丽红[2] 郑娟娟 胡海洋 WANG Wenyi;XIONG Jian;SONG Lihong;ZHENG Juanjuan;HU Haiyang(College of Computer and Information Engineering,Jiangxi Normal University,Nanchang Jiangxi 330022,China;Collge of Education,Jiangxi Normal University,Nanchang Jiangxi 330022,China)

机构地区：[1]江西师范大学计算机信息工程学院,江西南昌330022 [2]江西师范大学教育学院,江西南昌330022

出　　处：《江西师范大学学报（自然科学版）》2021年第2期118-125,共8页Journal of Jiangxi Normal University(Natural Science Edition)

基　　金：国家自然科学基金(62067005,61967009,31500909);汉考国际科研基金(CTI2019B10);江西师范大学教学改革研究(JXSDJG1848)资助项目.

摘　　要：大规模考试的公平性备受关注,探索能够生成多份平行测验并能保证测验质量的组卷方法十分重要.原用于计算机自适应测验(CAT)的基于最大优先级指标已开始用于组卷,但只局限于单题之间的匹配.为了增加求解空间,该文提出了一种基于最大优先级指标的双题匹配的组卷方法,实现题库中双题与种子试卷双题之间匹配,并结合平行试卷质量调整方法,用于生成平行试卷.将已有的组卷方法与双题匹配方法进行实验对比,结果表明:双题匹配方法有效地提升了生成的平行测验质量.The fairness of large-scale examinations has attracted much attention.It is very important to explore a method for generating multiple parallel tests and ensuring the quality of the tests.The maximum priority index(MPI)method originally proposed in the computer adaptive test(CAT)has been applied to constructing multiple parallel tests,but it is only limited to matching between single item and single item.In order to increase the solution space,a two items matching method based on the maximum priority index is proposed to realize the matching between the two items in the item bank and the two items in the seed test,and the parallel test quality adjustment method is combined to generate parallel tests.The existing test construction method is compared with the two items matching method.The results show that the two items matching method effectively improves the quality of the parallel test.

关 键 词：平行测验 组卷方法 计算机自适应测验 最大优先级指标 双题匹配 单题匹配 

分 类 号：B841[哲学宗教—基础心理学]