Study on the Enhancement of Personalized Borrowing Experience of Smart Library Users Based on Reinforcement Learning Framework

Abstract In the construction standard of smart library, personalized lending service enhances the important part of readers’ experience. The main idea of this paper is to achieve dynamic clustering of user groups through k-means and feedback, and a joint training method is proposed by combining backpropagation and reinforcement learning in order to compute the dynamic changes of user preferences, adjust the maximization gain strategy in time, and provide personalized borrowing content recommendation for users. The improved reinforcement learning method is used as a framework to build a personalized borrowing management system for smart library users, and the borrowing data and user preference features are converged and analyzed to provide support for the improvement of personalized borrowing experience. The results of the case analysis prove that the joint training method has the advantages of stability and fast convergence despite the increase in time consumption, while the algorithmic model in this paper has high recommendation accuracy and normalized discount cumulative gain. The personalized borrowing management system based on reinforcement learning effectively analyzes the user borrowing data and increases the number of borrowing by improving the borrowing experience, which shows that the work in this paper effectively improves the borrowing experience of readers.