基于在线学习行为的学习者画像构建
Construction of the Learner Portrait Based on Online Learning Behavior

学习者画像是大数据时代下的产物，其通过对学习者相关数据的分析来呈现学习者的学习特征，从而更好地为提升在线学习质量服务。如何挖掘分析学习者的相关数据来构建学习者画像是当前研究普遍关注的问题。本研究首先以日新学堂在线学习平台中的在线学习行为数据为切入点，基于交互视角对已有的在线学习行为数据进行分类，并据此划分学习者画像的维度。其次，在进一步明确学习者画像的标签后，通过采用K-means聚类方法形成三类学习者画像，并进一步将其命名为：高沉浸性学习者、中沉浸性学习者、低沉浸性学习者，对其画像进行描述，呈现三类学习者的在线学习行为特征。最后，提出学习预警是学习者画像的重要应用方向。
Learner portrait is a production in the era of big data. It presents the learning characteristics of learners through the analysis of learner-related data, so as to better serve to improve the quality of online learning. How to mine and analyze the relevant data of learners to construct learner portraits is a common concern of current research. This research takes the online learning behavior data in the Rixin Xuetang online learning platform as the starting point, and classifies the existing online learning behavior data based on the interactive perspective, and divides the dimensions of the learner portrait accordingly. After further clarifying the label of learner portraits, three types of learner portraits are formed by using k-means clustering method, and further named as high immersion learners, medium immersion learners and low immersion learners. Their portraits are described to present the learning behavior characteristics of the three types of learners. Finally, it is proposed that learning early warning is an important application direction of learner portrait.