Personalized Learning Resources Recommendation Based on Improved Attention Mechanism in Intelligent Online Education System

The current recommendation models in intelligent online education scenarios suffer from noise sensitivity and underutilization of graph relations, in order to compensate for these shortcomings, the study proposes a personalized learning resource recommendation model based on graph convolutional network with improved attention mechanism. The results show that when the noise category is feature noise, the detection time of the support vector machine combined with attention mechanism and the convolutional neural network combined with attention mechanism are 12.09s and 23.54s, respectively, while the detection time of the research model is only 5.25s. The detection accuracy of the research model is 98.37% when the data of learning resources is 3000, and the accuracy of the convolutional neural network combined with attention mechanism has the lowest accuracy of 74.24%. The above results show that the research model realizes the synergistic modeling of the structural features of the knowledge graph and the learner's behavioral sequences, and provides a reusable recommendation framework for the modern education system.