大数据环境下学业预警系统设计与分析

Design and Analysis of Learning Alert System in Big Data Condition

随着互联网+教育的深度融合，教育大数据逐渐成为学习分析的主要手段。该文讨论了构成教育大数 据环境的数据分类、收集与存储组织，以及教育数据挖掘和学习分析的一般思路与方法，探讨了在大数据环境下 的学业预警系统设计框架，提出了一种基于离群数据挖掘与分析的课程、课堂、课外“三位一体”预警信息发现 与生成模型LAOMA(Model of Learning Alert Based on Outlier Mining and Analysis)，建立了学业预警两类六级信号 系统及反馈机制。最后阐述了包括学业预警在内的学习分析技术使个性化学习真正成为可能，但如果应用不当， 过分依赖生硬的数据也可能与教育的本质不相符合。

Following the Internet deeply fusing with education， the educational big data are gradually taken for the principal environment in learning analysis. This paper discusses the classification， collection and storage organization of educational big data， and the general thought and method of educational data mining and learning analysis. A design framework of learning alert system in big data condition is presented in the paper. Further a model of Learning Alert based on Outlier Mining and Analysis (LAOMA) is put forward based on the trinity data of courses， classroom and outside class. And this model is used to discover and generate alert information. The two-type and six-level signal system of learning alert and its feedback mechanism are set up in this paper. Finally， the paper expounds that the individualized instruction can be realized by means of learning analysis technology inclusive learning alert， but inappropriate application of these technology and excessive dependent on rigid data are not likely to suit educational essence.