Modelling and trend analysis of student idea propagation paths facilitated by intelligent recommendation algorithms

This study develops a propagation path modelling method integrating implicit semantic analysis and ST-GCNs to examine how intelligent recommendation algorithms govern student idea dissemination. By constructing an algorithm-user-content dynamic interaction model, we quantify the impact of recommendation strategies on idea diffusion pathways within campus networks. Our analysis reveals a distinctive 'centralised-jumping' dual-mode evolution characterising the diffusion process. Validation using a publicly available social media dataset demonstrates a 12.7 percentage-point improvement in model accuracy over traditional methods. Furthermore, we find that educational level significantly alters path selection probability. This research provides educators and administrators with a theoretical framework and quantitative analytical tool for designing targeted information intervention strategies.