A Review of Learning Early Warning Research Based on Learner Portraits

In the context of big data in education, how to use the massive data of colleges and universities to better help students finish their studies on time and avoid the risk of not being able to finish their studies properly due to factors such as delay, retake, and drop out? The previous learning warning system built according to the grade ranking can no longer meet the requirements of today's universities. With the gradual maturity of data mining technology in education, building a learning alert system based on learner portraits has become a trend. By collecting and analyzing students' data and using data mining algorithms to accurately analyze their learning behaviors, the learner portrait provides a theoretical basis for learning early warning. This paper first takes learner portraits as a pavement, introduces learning early warning content using online learning, focuses on the research direction of learning early warning, then studies the current status of domestic and international research to summarize the specific algorithms of learner portraits and learning early warning applications, and finally summarizes the problems and presents the outlook. The innovation of the whole paper is to show the learner portrait through the visualization interface, and based on constructing the learner portrait, secondary data extraction is conducted to build the learning alert model for students with learning difficulties, which provides technical support for educators to implement personalized and accurate teaching.