Research on political Education Content Recommendation Model for Colleges and Universities Based on Graph Neural Networks

Abstract

As ideological and political education in colleges and universities enters the digital deepening stage, it becomes a key issue to build an intelligent recommendation system for the dual needs of semantic understanding and structural modeling. This paper proposes a recommendation model for ideological and political content that integrates knowledge graph construction and graph neural network modeling, generates multimodal knowledge graph through entity extraction and semantic relationship mining, and introduces graph convolution structure to jointly model content and user behavior. At the same time, the multimodal feature fusion and time-aware recommendation mechanism are designed to realize the dynamic portrayal of students' ideological interests and content adaptation.