【作 者】：
叶海智， 杨 柳， 黄宏涛， 梅钰皎
【关 键 词 】：
认知诊断； Rasch模型； 能力等级自适应； 试题推送模型； 个性化学习
【栏 目】：
学习环境与资源
【中文摘要】：
认知诊断与教育的深度融合能够促进个性化学习的发展。认知诊断测试中，学习者能力与测试题难度的匹配程度直接影响诊断结果的精确性。然而，目前认知诊断选题策略尚无法精确标定学习者能力及测试题目难度,导致教师不能准确掌握学生的学习状态。为解决该问题，构建了面向认知诊断的能力等级自适应试题推送模型，首先运用Rasch模型对题库原始数据进行对数转换，在同一等距量尺中标定出测试题目的难度等级和学习者的初始能力等级，在诊断测试过程中为不同能力等级的学习者推送相应难度等级的测试题目，然后在每次测试结束后更新学习者能力等级。实践结果表明：面向认知诊断的能力等级自适应试题推送模型能够使得测试题目难度与学习者能力精确匹配，进一步提高了认知诊断的精确性，提高了学生的学习效率，促进了教育向个性化方向的发展。
【英文摘要】：
The deep integration of cognitive diagnosis and education can promote the development of personalized learning. In the cognitive diagnostic test, the matching degree between the learner's ability and test difficulty directly affects the accuracy of diagnostic results. However, at present, the cognitive diagnosis cannot accurately calibrate the learner's ability and test difficulty, which leads to teachers' inability to accurately judge students' learning status. In order to solve that problem, a text push model adapted by the student's ability level for cognitive diagnosis is constructed. First, the Rasch model is used to perform logarithmic transformation on the raw data of the item bank, then the difficulty level of the test item and the learner's initial ability level are calibrated in the same isometric ruler. Learners with different ability levels are provided with corresponding test questions in the test， and the learner's ability level will be updated after each test. The result indicates that the text push model adapted by the student's ability level for cognitive diagnosis can make the difficulty of test items exactly match the learner's ability, further improve the accuracy of cognitive diagnosis, improves the learning efficiency of students and promotes the development of personalized education.