Behavior Analysis and Learning Outcomes Prediction Model Based on Blended Learning

Based on the related theories of learning behavior and previous studies, the characteristics of learning behavior in blended learning are discussed, and an analysis model of learning behavior in blended learning is established in this paper. Specifically, we first calculated the correlation between various learning behaviors and grades through the Pearson correlation coefficient; then, based on the k-means algorithm, we clustered the students, and explored the key factors that affect student performance through intra-cluster analysis; finally, we propose a student achievement prediction method based on a multi-layer perceptron machine, which accurately predicts a student’s achievement according to the student’s learning behavior. Through the above analysis results, some measures for blended instruction are put forward in this paper to improve students’ online learning behavior and enhance the blended learning outcomes.