融合遗忘行为和学习能力的深度知识追踪模型  

A deep knowledge tracing model integrating learning ability and forgetting behavior

作　　者：冯文芳[1] 刘杜奎 朱昶胜[2] FENG Wenfang;LIU Dukui;ZHU Changsheng(College of Economics and Management,Lanzhou University of Technology,Lanzhou 730050,China;College of Computer and Communication,Lanzhou University of Technology,Lanzhou 730050,China)

机构地区：[1]兰州理工大学经济管理学院,甘肃兰州730050 [2]兰州理工大学计算机与通信学院,甘肃兰州730050

出　　处：《武汉大学学报(工学版)》2025年第2期285-291,共7页Engineering Journal of Wuhan University

基　　金：甘肃省科技项目(编号:2014M560371);北京唐风汉语教育科技有限公司合作项目(编号:H2111cc004);兰州理工大学高等教育研究项目(编号:202102001)。

摘　　要：现有的知识追踪方法大多对遗忘因素建模不够健全或者忽略了学习者的学习能力,针对这些问题,提出一种融合遗忘行为和学习能力的深度知识追踪模型,该模型综合考虑了学习者的遗忘行为和学习能力,采用深度学习的方法结合遗忘因素对学习者的学习能力进行建模。在线教育数据集上的实验表明,相较于现有的知识追踪模型,融合遗忘行为和学习能力的深度知识追踪模型具有较好的可解释性和预测性,证明了融合遗忘因素和学习者学习能力的知识追踪模型的有效性。Most of the existing knowledge tracing methods are not sound enough to model forgetting factors,or ignore the learning ability of learners.To solve these problems,a deep knowledge tracing model integrating learning ability and forgetting behavior is proposed.This model comprehensively considers learners'forgetting behavior and learning ability,and adopts deep learning method combined with forgetting factors to model learners'learning ability.Experiments on online education datasets show that compared with existing knowledge tracing models,the deep knowledge tracking model integrating forgetting behavior and learning ability has better interpretability and predictability,which proves the effectiveness of the knowledge tracing model integrating forgetting factors and learners'learning ability.

关 键 词：知识追踪 深度神经网络 遗忘行为 学习能力 在线学习 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]