Research on Recommendation Model of Online Educational Resources Based on Compressed Interaction and Feature Weighting

Abstract

The construction of smart cities provides broad application scenarios and rich data resources for smart education. With the aid of cutting-edge technologies like the Internet of Things, big data, and cloud computing in the context of smart cities, smart education can achieve more accurate teaching management, personalized learning services, and intelligent educational evaluation. In the context of smart cities, educational informatization and educational modernization have received technical support, and artificial intelligence technology represented by recommendation algorithms has attracted widespread attention from academia and industry. Based on the real data resources of educational platform, combined with the characteristics of implicit interaction data mining, aiming to address the deficiencies of current recommendation algorithms in exploiting implicit interaction data, this paper puts forward a recommendation model grounded on the Compress Interaction and Feature Weight Network (CIFW). The model uses channel attention mechanism and bilinear feature interaction to strengthen implicit feature interaction mining. Considering the limitation of implicit interaction in interpretability, we introduce Compress Interaction Network (CIN) as explicit feature interaction to improve the interpretability of the model. Moreover, through the integration of nonlinear and linear outcomes, the model's generalization and memory capabilities are bolstered. Comparative experiments indicate that, in terms of accuracy and the area under the ROC curve, the CIFW model proposed in this paper outperforms other commonly utilized recommendation models.