中文摘要：

在线学习者学习投入度是影响在线学习效果、满意度和质量的重要因素,是远程教育和在线教育领域研究的热点之一。文章针对在线学习中的独立自主学习,根据x API标准所提出的在线学习活动,确定当前能反应独立自主学习活动的三类数据——教学视频播放行为数据、视频观看时长和并发学习行为数据,提出在线学习投入度算法,实现了对自主学习投入度的实时观察及计算。基于提出的自主学习投入度算法,对1万多在线学习者真实的在线学习行为数据进行挖掘和分析,主要研究结论包括：（1）实时了解和监控自主学习过程中群体和个体投入情况的特点,可以促使学习者准确知晓并自我调整学习状态,实现教学方的有效监控;（2）通过分析投入度数据,能够挖掘一些难以直接观察到的在线自主学习规律和特点;（3）有必要动态分析视频学习材料的学习投入度情况,从而辨析视频学习材料的有效性和可行性,这能为视频学习材料的设计、制作、调整和修改提供数据支持和依据,避免设计制作的盲目性和主观性。

英文摘要：

Online learners＇ learning engagement is one of important factors influencing the effectiveness, satisfaction and quality of online learning. It is a hot issue in distance education and online education. For independent learning of online learning, according to the online learning activities based on x API standard, this paper selects three types of data, namely teaching video play behavior data and video viewing time and concurrent learning behavior data. An online learning engagement algorithm is proposed to realize the real-time observation and calculation of learning engagement. The autonomous learning engagement algorithm is used to deal with the data mining and analysis of more than 10 thousand online learners＇ authentic online learning behaviors. The main results include：（1） the real-time understanding and monitoring of the characteristics of the group and individual engagement in the process of autonomous learning can help learners to know and adjust their learning status accurately and realize effective monitoring of the teaching side;（2） Through the analysis of engagement data, rules and characteristics difficult to be observed directly of online independent learning can be found out;（3） It is necessary to dynamically analyze learners＇ learning engagement of video learning materials in order to differentiate the effectiveness and feasibility of those materials, which can provide data support and basis for the design,manufacture, adjustment and modification of video learning materials and avoid the blindness and subjectivity in design and manufacture.