A Machine Learning-Based Design Case Study on the Influencing Factors of Classroom Silence Among Chinese University Students: The “Reconciliation Classroom” Conceptual Healing Space

Classroom interaction among students is an externalization and expression of knowledge, playing a crucial role in facilitating learning. However, university students in China tend to remain silent in class, rarely initiating or responding to questions. This study investigates the phenomenon of classroom silence among university students by employing quantitative statistical methods to explore the influence of classroom-related factors on students’ level of classroom engagement. A total of 449 questionnaires were distributed and collected, and after verifying the reliability and validity of the questionnaire, data cleaning, and the exclusion of outliers, 357 valid samples were obtained for multiple regression and machine learning analyses. Results show that the multiple regression model performs better in predicting the independent variables and the random forest (RF) model outperforms in predicting the most influencing factor on the classroom interactive activity level. Performance accuracy achieved by the multiple regression model is 100% compared to the RF model with 80% accuracy. The findings indicate that (1) students’ personality traits exhibit a significant positive correlation with students' classroom engagement; (2) students' abilities have a notable positive impact on their participation in classroom interaction; and (3) teachers’ teaching styles and response benefits demonstrate a weak positive correlation with classroom engagement. Based on the implementation of machine learning models, a new idea emerges to tackle students’ silence in classrooms by proposing and implementing a predictive modeling approach to identify students at risk in learning in universities and colleges.