基于多特征提取的时间卷积知识追踪模型  

A Temporal Convolutional Knowledge Tracking Model Based on Multiple Feature Extraction

作　　者：谢仕斌 刘梦赤 唐诗琪 周瑞平 XIE Shi-bin;LIU Meng-chi;TANG Shi-qi;ZHOU Rui-ping(School of Computer Science,South China Normal University,Guangzhou 510631,China;Guangzhou Key Laboratory of Big Data and Intelligent Education,South China Normal University,Guangzhou 510631,China)

机构地区：[1]华南师范大学计算机学院,广东广州510631 [2]华南师范大学广州大数据智能教育重点实验室,广东广州510631

出　　处：《计算机与现代化》2023年第7期25-29,42,共6页Computer and Modernization

基　　金：国家自然科学基金资助项目(61672389);广州市大数据智能教育重点实验室项目(201905010009)。

摘　　要：知识追踪(Knowledge Tracing,KT)是教育数据挖掘领域中的关键技术,其通过利用学生的历史学习记录来预测学生下一次的作答表现。针对基于时间卷积网络(TCN)的深度知识追踪模型存在的只使用学生答题序列和答题结果,而忽略学生其他行为特征的问题,本文提出一种基于多特征提取的时间卷积知识追踪模型TKT-PCA。使用时间卷积网络和主成分分析(Principal Component Analysis,PCA)方法自动提取多种学生作答行为中的隐藏特征并学习它们的表示,不仅降低了特征维度减少冗余信息,还充分评估了学生的知识掌握情况。实验结果表明,与其他知识追踪基线模型相比,TKT-PCA有更好的预测性能。Knowledge tracing(KT)is a key technology in the field of educational data mining.It uses students’historical learn⁃ing records to predict students’next answer performance.Aiming at the problem that the deep knowledge tracking model based on time convolution network(TCN)only uses students’answer sequences and answer results,and ignores other behavior charac⁃teristics of students,a deep knowledge tracking model based on multi feature extraction(TKT-PCA)is proposed.The model uses principal component analysis(PCA)method to automatically extract hidden features in a variety of students’answer behav⁃ior and learn their representation.It not only reduces the feature dimension and redundant information,but also fully evaluates students’knowledge mastery.The experimental results show that the TKT-PCA has the better prediction performance compared with other knowledge tracking baseline models.

关 键 词：深度学习 知识追踪 时间卷积网络 教育数据挖掘 智能教育 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]