基于LSTM深度学习的在线学习预测模型研究

Research on Online Learning Prediction Model Based on LSTM Deep Learning

分析大规模开放在线课程(MOOC)在线学习学生的课程学习表现有助于教师及时进行干预,提高教育教学质量。为了能在课程早期预测学生的表现,本文建立一个基于长短期记忆(LSTM)深度学习的N-LSTM模型,通过使用学生基本信息和活动点击流数据,在课程开始的初期即可周期性预测MOOC中4类学生表现(退学、不及格、通过和杰出)。为了检验该模型的有效性,将其与循环神经网络(RNN)和门控循环神经网络(GRU)进行了比较,实验结果表明,N-LSTM模型准确率为72%,优于RNN模型63%的准确率和GRU模型66%的准确率,效果最好。

Analyzing the course learning performance of MOOC online learning students can help teachers intervene in a timely manner and improve the quality of education and teaching.In order to predict student performance early in the course,this study establishes a new model based on LSTM deep learning.By using student basic information and activity click stream data,the performance of four types of students(dropout,fail,pass,and outstanding)in open online courses(MOOC)can be periodically predicted at the beginning of the course.In order to test the effectiveness of this model,it is compared with recurrent neural network(RNN)and gated recurrent neural network(GRU).The experimental results show that the accuracy of the model proposed in this study is 72%,which is better than the accuracy of RNN model by 63%and GRU model by 66%.