This research proposes a novel recommendation model that enhances career planning accuracy for college students, implying improved guidance for their future.
Key Points
The aim is to improve the accuracy of career planning recommendations for college students.
Constructed a comprehensive career planning knowledge graph.
Projected sparse user and item vectors into a dense latent space.
Employed a Graph Convolutional Network to aggregate neighbor vectors.
Integrated graph-based representations with a matrix factorization component.
Utilized a fully connected layer for final recommendation scoring.
Achieved recommendation accuracy improvements of at least 2.04%, 2.26%, and 1.34% over existing methods on real-world datasets.