摘 要: 混合式教学的普及使得电子作业成为一种评价学生学习效果的重要数据来源,利用机器学习对电子作业 进行建模是对学生学习预警的一种有益探索。本文在对电子作业进行分词和向量化基础上,通过k-means聚类和轮廓系 数来判断其语义的多样性,通过计算文档向量相似性矩阵的网络效率来评价电子作业的中心性。实验结果显示,该方法 可以有效寻找电子作业聚类效果最优时的簇类多样性,也可以有效评价电子作业相似度的网络中心性。因此,该方法作 为一种学生学习预警模型,可以对电子作业文档的多样性和中心性给出客观的总体评价。 |
关键词: 文档向量;k-means聚类;轮廓系数;文档相似度;图论效率 |
中图分类号: TP181
文献标识码: A
|
基金项目: 本文系江苏科技大学2016高教研究立项课题资助项目“基于电子作业挖掘的学生学习预警模型研究”(项目编号:GJKTY201625);本文系江苏科技大学2015年学校重 点教改课题资助项目“计算机类专业通用课程优质教学资源建设的研究与实践”;本文系教育部在线教学研究中心2017混合教学试点单位项目(项目编号:2017137);本 文系江苏省教育信息化研究课题资助项目“基于云计算的泛在学习生态系统研究与实现”(项目编号:20172217). |
|
Research on School Precaution Model Based on Electronic Assignments Mining |
ZHANG Xiaofei,DUAN Xianhua,LIU Zhen,QIAN Ping
|
( School of Computer, Jiangsu University of Science and Technology, Zhenjiang 212003, China)
|
Abstract: The popularity of hybrid teaching makes electronic assignment an important data source for evaluating students' learning effects.Modeling electronic assignment with machine learning is a useful exploration for school precaution. Based on the word segmentation and vectorization of electronic assignments,this paper determines the semantic diversity by k-means clustering and silhouette coefficient,and evaluates the centrality of electronic assignment by calculating the network efficiency of document vector similarity matrix.The experimental results show that the method can effectively find the cluster diversity when the clustering effect of electronic assignments is optimal,and can also effectively evaluate the network centrality of the similarities of electronic assignments.Therefore,as a school precaution model,this method can give an objective overall evaluation of the diversity and centrality of electronic assignments. |
Keywords: document vector;k-means clustering;Silhouette coefficient;documents' similarity;graph theoretic efficiency |