基于KNN算法的mCSCL学习伙伴分组策略研究

mCSCL Learning Partner Grouping Strategy Research Based on KNN Clustering Algorithm

随着信息技术的快速发展,mCSCL已成为教育技术学领域新的研究热点,学习伙伴选择合理与否将直接影响着协作学习效率。文章利用mCSCL环境下协作分组伙伴模型,提出了基于KNN的mCSCL学习伙伴分组理论,通过计算学习者之间的相似度和类别权重,提供一张可视化的学习伙伴关系图,导学者遵循组间同质和组内异质分组原则,为学习者动态推荐最佳学习伙伴;并设计了以小学一年级加减运算为内容的mCSCL活动,开展分组满意度访谈和小组学习效率实证研究。实验结果表明,相对于随机分组方式,基于KNN算法的mCSCL学习伙伴分组方式更适合移动学习活动开展,学习效率更高。

With the rapid development of information technology, mCSCL has become a new field of educational technology research focused, whether is reasonable selection of learning partner will direct impact the efficiency of learning. This article uses the collaboration partner grouping model in mCSCL environment and proposes a mCSCL learning partner grouping theory based on KNN algorithm, which can provide a Visual Learning Partnership map by computing the similarity between learners and classification weight. The scholars recommend the best learning partner for learner dynamically according to the grouping principles of homogeneous between group and heterogeneous in the group to. Paper designed a mCSCL activities about the first grade of addition and subtraction, developed the interview of the satisfaction about the grouping of learning partner and the empirical study about the learning group＇s study efficiency. Experimental results show that the way of mCSCL learning partner grouping based on KNN algorithm is more appropriate for mobile learning activities ,and learning more efficient relative to the random grouping.