北京工业职业技术学院学报

为解决传统学情数据的维度单一与干预滞后等问题，采用数据驱动与教育理论融合的分析框架，设计了基于动态多维学情画像的学生学业预警与干预系统。系统通过整合水平(历史成绩)、过程(学习投入)和趋势(学业风险)3类数据，构建实时更新的学情画像模型，实现了学业风险的实时预警与针对干预；建立“数据聚合—画像构建—预警干预—效果反馈”的完整闭环，实现了从静态评估向动态追踪的转变。应用结果表明：系统能将被干预者平均成绩提升29%;能有效识别群体学习薄弱点；学情分析效率提高95%,在提升分析效能与干预及时性上表现良好。

To solve the issues of one-dimensionality and intervention lag in traditional learning analytics data, a data-driven and educational theory integrated analysis framework was adopted to design a student academic warning and intervention system based on dynamic multidimensional learning profiles. By incorporating three categories of data, level(historical academic performance), process(learning engagement), and trend(academic risk), the system builds a real-time updated learning profile model, enabling real-time early warning and targeted intervention for academic risks. A complete closed-loop workflow of “data aggregation-profile construction-early warning and intervention-effect feedback” was established, which achieves the transformation from static assessment to dynamic tracking. Application results demonstrate that the system increases the average score of intervened students by 29%, effectively identifies collective learning weaknesses, and improves the efficiency of learning analytics by 95%, thus exhibiting excellent performance in enhancing analytical effectiveness and intervention timeliness.