Factors influencing peer interaction among college students in blended learning environments: a study based on SEM and ANN

In actual blended learning environments, the quality and depth of peer interaction still face many challenges. The current research on the factors influencing peer interaction is not comprehensive, particularly lacking systematic analysis on how to improve the level of peer interaction. Based on Social Learning Theory and the Community of Inquiry (CoI) framework, this study constructs a hypothetical model to explore the key factors affecting university students' peer interaction. Structural Equation Modeling (SEM) was used to analyze 300 questionnaire samples, indicating that learning motivation, personality traits, diverse interaction environments, learning tasks, grouping methods, and teacher support significantly influence peer interaction effectiveness. Complementary Artificial Neural Networks analysis shows that grouping methods is the most important factor in predicting peer interaction, followed by personality traits, interaction environment, learning tasks, and learning motivation. Based on these findings, the study proposes several strategies to enhance peer interaction levels, including self-paced learning strategies based on micro-videos, collaborative learning strategies based on heterogeneous grouping, teacher-student assistance strategies based on Blended Learning Environment, and peer review strategies based on self-reflection. This research provides valuable insights into optimizing peer interaction in blended learning, contributing to the development of more effective educational practices.