摘要
聚类统计分析在大数据理论研究和实践应用方面具有重要地位,是学习分析技术的重要组成部分。文章首先在数据初始化和规范化的基础上定位分类条件,实现学习行为的分化和集成,形成多个待聚类的数据子集;然后,根据学习交互活动之间的拓扑关联性和依赖性,设计随机游走模型与BIRCH算法融合的聚类统计方法,实现关键学习交互活动的检索评估和数据聚类;最后,对算法执行的多个性能指标进行计算和对比。实验结果表明,改进后的算法在学习交互活动聚类方面具有明显优势,聚类统计过程和分析结果具有可行性和可靠性。
Clustering statistical analysis plays an important role in the theoretical researches and application of education big data,and it is also an important part of learning analytics.On the basis of data initialization and standardization,this paper firstly locates the classification conditions,and realizes the differentiation and integration of learning behavior,forming multiple data subsets to be clustered.Then,according to the topological correlation and dependence between interactive learning activities,the paper designs the clustering statistical method combining random walk model and BIRCH algorithm to realize the retrieval evaluation and data clustering for key learning interactions,and finally calculates and compares the multiple performance indicators performed by the algorithm.The results of experiments show that the improved algorithm has obvious advantages in clustering of learning interactions,and that the clustering statistical process and analysis results are feasible and reliable.
作者
夏小娜
Xia Xiaona(Faculty of Education,Qufu Normal University,Qufu Shandong 273165,China;Chinese Academy of Education Big Data,Qufu Normal University,Qufu Shandong 273165,China)
出处
《统计与决策》
CSSCI
北大核心
2021年第23期5-9,共5页
Statistics & Decision
基金
山东省教育科学重点课题(2020ZD030)。