Abstract:  Evaluation is the analysis of performance and strategy effectiveness, and it is an important dimension of metacognition. Evaluation can promote learners to implement self-regulation to adapt to the autonomous learning process. Learning analytics is used to establish the relationship between learning behaviors and evaluation ability, in order to explore the feasibility of using behavioral data to predict the development of evaluation ability. This study took students in the professional introduction course of School of Information and Communication Engineering, Beijing University of Posts and Telecommunications as the research object, used evaluation scale to assess the evaluation level of the students to analyze their evaluation ability’s changes, and constructed learning behavioral indicators based on the course learning process. After selecting learning behaviors and processing and analyzing prediction models, the results show that the four types of learning behaviors: resource utilization, self-checking, review and task completion quality can predict the changes of evaluation ability, and demonstrate the feasibility of predicting the development of metacognitive evaluation ability based on online learning behavioral data. This study provides reference for teachers to train students’ evaluation ability and improve their autonomous learning ability.

Key words:  , learning behavior, evaluation, prediction, metacognition, autonomous learning