基于循环门单元和注意力机制的学生学习积极性预测模型

Prediction model of students’ learning enthusiasm based on gated recurrent unit-attention

为了挖掘校园无线网络数据中蕴含的学生行为,更好地辅助教学管理,本文提出了基于循环门单元和注意力机制(gated recurrent unit-attention,GRU-Attention)的学生学习积极性预测模型。首先对采集到的学生行为轨迹数据进行预处理和聚类分析,根据学生在校园各个场所的平均逗留时间对学习积极性做出预判断。然后根据无线流量的连接频率及用途对学生行为特征数据进行二次判断,并引入学生学习成绩作为参考标准。最后将数据集输入到GRU信息提取层,并在Attention层进行加权处理,计算得出概率分布。实验结果表明,GRU-Attention模型准确度优于其他常见预测模型,可以更好地帮助辅导员去管理和引导学生树立良好的学习态度。

In order to mine the students’ behavior contained in the campus wireless network data and better assist the teaching management, this paper proposes a prediction model of students’ learning enthusiasm based on gated recurrent unit-attention(GRU-Attention). Firstly, the collected data of students’ behavior trajectory are preprocessed and cluster analyzed, and the learning enthusiasm is pre judged according to the average stay time of students in various places on campus. Then, according to the connection frequency and usage of wireless traffic, the data of students’ behavior characteristics are judged twice, and the students’ academic performance is introduced as the reference standard. Finally,the data set is input into the GRU information extraction layer, and weighted in the attention layer to calculate the probability distribution. The experimental results show that the accuracy of GRU-Attention model is better than other common prediction models, which can help counselors guide students with low enthusiasm.