Intelligent recommendation and optimization system of educational resources based on DNN

Abstract

This paper mainly discusses the application and optimization of deep learning neural networks in intelligent recommendation systems of educational resources. The research first introduces the key algorithm models and related concepts in recommendation algorithms. It includes the basic principles of deep learning models and their specific applications in the intelligent recommendation of educational resources based on DNN. This paper proposes a recommendation algorithm model combined with Gradient Boosting DeepFM regarding system algorithm construction. The model integrates the gradient lifting decision tree module and the factorization machine module, aiming to improve the recommendation system’s accuracy and personalization. Among them, the gradient lifting decision tree module iteratively trains multiple weak classifiers and gradually approaches the optimal solution, accurately capturing users’ learning behaviors and preferences. The factorization machine module effectively improves the generalization ability of the recommendation system by modeling the potential characteristics of users and resources. In the experimental part, the study was tested on a real dataset containing 10,000 users and 50,000 educational resources. The experimental results show that the recommendation system based on Gradient Boosting DeepFM improves by 15%, 12%, and 13.5% in accuracy, recall, and F1 score, respectively, compared with traditional recommendation algorithms. Especially in terms of user satisfaction, the degree of matching between the educational resources recommended by the system and the actual needs of users is as high as 85%, which is 20 percentage points higher than the traditional method. The intelligent recommendation and optimization system of educational resources based on DNN proposed in this study effectively improves the accuracy and personalization of educational resource recommendation through optimizing algorithm models and experimental verification. It provides strong technical support for online education platforms.