不定长CAT引入曝光因子的平均调整信息选题策略研究  

New Item Selection Strategy of Combination of Exposure-control Factor with Average Adjustment Information for Variable Length Computerized Adaptive Testing

作　　者：朱隆尹[1,2] 丁树良[1] 程小扬[1,2] 蓝立毅[1,2] 

机构地区：[1]江西师范大学,南昌330022 [2]赣南师范学院,赣州341000

出　　处：《心理学探新》2015年第1期68-71,共4页Psychological Exploration

基　　金：国家自然科学基金(60263005;30860084;31160203;31360237);江西省教育厅科技项目基金(GJJ10238;GJJ13206;GJJ13208;GJJ13209;GJJ13227)

摘　　要：选题策略是CAT关键技术之一,好的选题策略应该能够较好地权衡考生能力估计精度和测验的安全性。从引入曝光因子的选题策略入手,通过调整信息平均的方法提出了引入曝光因子的平均调整信息选题法,实验数据表明新的选题策略能大大增强测验的安全性而不降低能力估计的精度。Item selection strategy is one of the crucial techniques of computerized adaptive testing. Both the accuracy of ability estimation and the security of test will be greatly improved with appropriate item selection strategy. This paper proposes an average adjustment information strategy through introducing exposure - control factor and average adjustment information method in the item selection. The experiment result shows that the security of the test will be enormously enhanced by the new item selection strategy.

关 键 词：不定长CAT选题策略 曝光因子 极大平均调整信息策略 优化的极大平均调整信息策略 

分 类 号：B841.2[哲学宗教—基础心理学]