摘要
学习者的个体特征和学习过程中的行为表现对在线学习结果具有一定的预测作用。为了探究在职教师在线学习结果的影响因素及预测模型,研究以教师信息技术应用能力在线学习为研究场景,选取1041名教师作为研究对象,依据性别、年龄、学科、学段等倾向性指标和参与、专注、绩效、规律4个维度的行为表现指标建立4种学习结果预测模型,综合评估7种分类算法后选择较优的预测模型和算法,并对确定的预测模型进行调参优化、动态适应性检验、可视化呈现和规则提取,并对预测风险给出了干预策略。研究表明:应用CART决策树算法通过倾向性指标和行为表现指标的混合预测模型获得了较优的预测效果,并且该模型具有早期的预测能力,可以为培训管理者在不同的学习阶段实施学习干预和支持服务提供科学依据。
Online learning results can be predicted through learners’individual characteristics and behavior in the learning process.In order to explore the influencing factors and prediction model of learning results in Teachers’online professional development,this paper takes the project of teachers’information technology application ability online learning as the research scene,and 1041 teachers are taken as the research object.The four learning results prediction models are established according to gender,age,discipline,school stage and other dispositional indicators and behavior indicators such as participation,attention,performance and regularity.After seven classification algorithms are evaluated,the better prediction model and algorithm are selected.Then the generalization ability of the model is optimized by adjusting parameters,the dynamic adaptability is tested,and the knowledge and rules are visualized and extracted from the model.Finally,some intervention strategies for predicting risk are given.The results show that better prediction performance are obtained by using the classification and regression trees(CART)algorithm in the hybrid prediction model of dispositional indicators and behavior indicators.The model has early prediction ability and can provide scientific suggestions for training managers to implement learning intervention and support services in different learning stages.
作者
李昕
荆永君
Li Xin;Jing Yongjun(Institute of Educational Technology,Shenyang Normal University,Shenyang 110034,Liaoning)
出处
《中国电化教育》
CSSCI
北大核心
2022年第6期96-104,共9页
China Educational Technology
基金
全国教育科学规划教育部重点课题“基于学习分析的教师网络学习行为预测与干预研究”(课题编号:DCA170305)研究成果。
关键词
在线专业发展
倾向性指标
学习行为
学习结果
预测与干预
online professional development
dispositional indicators
learning behavior
learning results
prediction and intervention
