Optimization method for educational resource recommendation combining LSTM and feature weighting

With the continuous development of the Internet and the growing demand for education, various online courses have emerged in large numbers. Learners are facing resource overload and screening difficulties. Therefore, classifying and recommending educational resources has become an essential task for various platforms. Ordinary educational resource recommendation models are usually based on simple search functions and user profiles for recommendation. Faced with massive educational resources and low user interaction behavior, basic recommendation methods have low efficiency and serious information loss. Therefore, a personalized educational resource recommendation model on the basis of bidirectional long short-term memory and feature weighting is proposed in the experiment. The compressed interactive network is used to clarify the data, and the channel attention mechanism is used to weight the data. Finally, the bidirectional long short-term memory network algorithm is used for encoding iteration to minimize data omission and improve data interactivity, achieving accurate recommendation of educational resources. The goal of the research is to divide large-scale resources into specific disciplines, types, and precise resources with user attributes based on the characteristics of educational resource data and user related interactive behaviors. The constructed model exhibited a loss function value below 0.4, a response time of less than 400ms, and a recommendation accuracy of over 80% on the relevant dataset. In practical applications, the accuracy of classifying educational resource data on Xuetang Online website and Tencent Classroom reached 0.99, and the recommendation satisfaction rate reached 0.8. The above data indicates that the personalized education resource recommendation model has good classification recommendation accuracy and stability, which can participate well in personalized education resource recommendation work, achieving optimization and upgrading of education resource recommendation methods on different platforms.