基于遗传算法的组合自适应考试模型与方法_白岩.pdf

2011 International Conference on Education and Education Management   A Genetic Algorithm-based Assembled Computerized Adaptive Testing Model and Method Yan bai1, a,Zhong Yao2, b and Jinxin Hao3, c 1, 2, 3Beihang University School of Economics and Management，Beijing，China a b c xiaoyoulang@ ， iszhyao@， gordonhao@ Keywords:Classical test theory, item response theory, genetic algorithm,adaptive test Abstract. In this paper, we proposed an assembled computerized adaptive testing model and method by analyzing the advantages and disadvantages of Classical Test Theory and Item Response Theory. By differentiating the medium-level examinees with high-level and low-level examinees, the model begins with generating exam paper based on Genetic Algorithm, which followed by roughly evaluating examinees’ ability. For medium-level examinees, the model will directly use the generated exam paper; while for the high-level and low-level examinees, the model will dynamically adjust some parts of the exam paper under the constraints of knowledge coverage distribution of the generated exam paper. Based on the model, related method is also proposed and implemented on Moodle course management system. 基于遗传算法的组合自适应考试模型与方法 1, a 2, b 3, c 白岩 ，姚忠 ,郝金星 1, 2, 3北京航空航天大学经济管理学院，北京，中国 a b c xiaoyoulang@ ， iszhyao@， gordonhao@ 关键词：经典测试理论；项目反应理论；遗传算法；自适应考试 中文摘要.在分析经典测验理论和项目反应理论优缺点的基础上，提出了一种组合自