Analysis of online learning behavior for the course of computer composition and architecture

In order to establish layered learning profiles of students in the course on Computer Organization and Architecture, data for online learning are collected on eight dimensions, which include task completion rate, chapter study count, video viewing duration, chapter quiz completion rate, course points, checked-in frequency, in-class exercise count and exam score. Firstly, the K-means++ algorithm is employed to perform cluster analysis on different student groups. Subsequently, the heatmap is utilized to visually evaluate the correlation between any two learning dimensions. Finally, based on the results of data analysis, suggestions for teaching improvement are provided.