认知诊断支持下的合作学习互补分组方法研究

分组是合作学习的重要环节,能够培养学生的小组协作能力,提高学生合作学习的效率.然而,目前合作学习常采用人工方式进行分组,无法确保组内成员在知识结构上具有较高的互补性.因此,提出一种基于认知诊断的分组方法.该方法首先使用多层感知器神经网络对学生知识状态进行诊断,在诊断结果支持下通过获取知识点间的依赖关系生成知识状态迁移图并对其进行拓扑排序,最后按知识状态拓扑序对学生进行自动分组,使分组内成员在知识结构上具有较高的互补性.在此基础上应用该分组方法进行合作学习教学设计,并将其应用于《C程序设计》课程中开展教学实践.实践结果表明:该方法能够有效提高小组成员在知识结构上的互补性,从而提高问题解决率,改善学生学习效果.问卷调查结果显示:参与实验的学生对基于互补分组的教学方式满意度较高.

Research on complementary grouping method for cooperative learning supported by cognitive diagnosis

Grouping is an important part of cooperative learning,which can build collaboration ability among group members and improve their efficiency of cooperative learning.However,nowadays cooperative learning is normally divided into groups by hand,which can't ensure the highly complementary about the knowledge structure of group members.Therefore,the paper proposes a complementary grouping method based on cognitive diagnosis.Firstly we uses uses a multi-layer perceptron neural network to diagnose students' knowledge states.Under the support of diagnosis results,the knowledge state transition diagram is generated and topologically sorted by obtaining the dependencies among knowledge points.Then the students are automatically grouped according to the topological order of the knowledge state,so that the internal members of the group have a relatively high complementarity in the knowledge structure.This grouping method is used to carry out instructional design based on cooperative learning and teaching practice in C Programming.The experimental data show that this method can effectively improve the complementarity of the knowledge structure among the members of the group,thereby increasing the problem solving rate and improving student learning.Questionnaire survey results showed that students participating in the experiment were more satisfied with the teaching methods based on complementary groupings.