Personalized Learning Resource Recommendation Algorithm Based on Autoencoder

With the development of information technology and the popularization of the Internet, personalized learning resource recommendation plays an increasingly important role in improving the quality and efficiency of education. Personalized learning can provide customized learning content according to learners' interests, knowledge level and learning habits, so as to stimulate learners' learning motivation and improve learning results. However, traditional recommendation algorithms such as collaborative filtering and content-based methods face challenges in dealing with sparse data and cold start problems, and it is difficult to fully meet the needs of personalized learning. In this paper, a personalized learning resource recommendation algorithm (Aera) based on Autoencoder is proposed. The algorithm combines the advantages of collaborative filtering and traditional content-based recommendation methods, and further improves the accuracy and personalization of recommendations through deep learning technology. As an unsupervised learning model, autoencoder can effectively extract key features from high-dimensional sparse data and solve the problem of data sparsity in traditional recommendation algorithms. In addition, through deep feature mining of user behavior data, this algorithm can more accurately capture the changes of user interest and provide users with more personalized learning resource recommendation. The experimental results show that compared with the traditional recommendation methods, the auto-encoder based recommendation algorithm not only improves the accuracy of recommendation, but also enhances the user experience, which is of great significance for promoting personalized learning.