【摘要】 有效分组是提升协作学习效率的关键，合理的分组能够使个体和组内成员均获得最大化的习得成效；然而，目前的分组方法未对学习者知识水平特征深度计算，无法保证组内成员在知识结构上相辅相成。因此，提出一种融入深度知识追踪优化模型(DKVMN-KT)的协作学习分组方法。首先，采用深度知识追踪优化模型对学习者的知识状态建模，得到学习者的知识掌握程度；然后，利用K-means方法对所有学习者进行相似聚类，最后根据分组的异质性原则，将不同簇的学习者分配到适合的学习小组。实验结果表明：该方法能够实现学习者知识结构层面的有效分组，分组结果在知识结构上更具有公平性和异质性。更多还原

【Abstract】 Effective grouping in collaborative learning is a key factor for improving the efficiency of collaborative learning.The reasonable grouping can maximize the learning outcomes for both individuals and members within the group. However, the current cooperative learning grouping method lacks deep calculation on the characteristics of the knowledge level of learners, and it is impossible to ensure that members in the group complement each other in the knowledge structure. Therefore, a collaborative learning grouping method incorporating the optimization strategy of Deep knowledge tracking(DKVMN-KT) is proposed. Firstly, the optimized DKVMN-KT model is used to model the knowledge state of learners to obtain their level of knowledge mastery. Then the K-means method is used to similarly cluster on the mastery levels of all learners. Finally, the learners of different clusters are assigned to suitable learning groups according to the heterogeneity principle of grouping. The experimental results show that this approach achieves effective grouping at the level of the learners’ knowledge structure. The grouping results are more fair and heterogeneous in terms of the knowledge structure.更多还原