基于Transformer模型的学情预警系统

A School Precaution System Based on Transformer Model

学情预警是学生管理工作的重要组成,是日常教学管理和教学改革的重要参考。目前学情预警主要面临学情评价不全面、预警不及时的问题。本文提出了一种基于Transformer模型的学情预警系统,分为数据存储、预测系统和可视化界面三个部分。数据存储通过ETL工具将相关数据提取至数据仓库;预测系统使用Transformer模型根据学生历史数据和风险指数推测本学期课程成绩;可视化界面将预警结果分层并展示。经实验验证,系统准确性能够满足使用需求,使相关人员更及时更全面了解学生的学习状态,开展相关工作。

Learning situation warning is an important component of student management work and an important reference for daily teaching management and teaching reform.At present,the main problem facing the early warning of academic situation is the incomplete evaluation of academic situation and the untimely warning.This article proposes a learning situation early warning system based on the Transformer model,which is divided into three parts:data storage,prediction system,and visualization interface.Data storage extracts relevant data into a data warehouse through ETL tools;The prediction system uses the Transformer model to predict the course grades of this semester based on student historical data and risk index;The visualization interface layers and displays the warning results.Through experimental verification,the accuracy of the system can meet the usage needs,enabling relevant personnel to timely and comprehensively understand the learning status of students and carry out related work.