学习分析作为学习科学的子领域,其关注的核心问题就是对学习过程的理解与优化,而这离不开对学习者学习行为数据的收集和对行为模式的分析.特征工程作为一种基于底层数据设计特征集的系统方法,为行为模式的分析提供了新的技术支持与研究路径.文章通过介绍特征工程实施的四个步骤,系统梳理了目前使用特征工程方法识别出的典型行为模式,如投机取巧、挫折、疑惑等,可为行为模式的相关研究提供参考.同时,文章基于对有效的技术支持和实践意义两个话题的讨论,指出了未来的研究取径与研究重点.%As a subfield of Learning Science, the key issue of learning analysis is to understand and optimize learning and the environment, which is significantly associated with data collection of learning behaviors and analysis of behavior patterns. As a method designing the feature set based on the underlying data, the feature engineering system provides new technical support and research path for the analysis of behavior patterns. This paper introduces the fundamental steps of feature engineering, and further systemically reviews some typical types of behavior patterns distinguished by feature engineering, such as gaming the system, frustration and confusion. Finally, this paper hopes to provide references for future studies of behavior patterns. Meanwhile, it points out the possible research direction and research focuses by discussing the effective technical support and practical significance.
展开▼
