Portal de Periódicos da CAPES

Strategies for Improving Collaborative Filtering Library Circulation Services Based on AR Technology Under the Perspective of Five Education Initiatives

2024; Springer Nature; Linguagem: Inglês

10.1007/978-981-99-6956-2_17

2190-3026

Zhenwei Wang,

Robotics and Automated Systems

In order to solve the problem of difficult internal guidance in university libraries, a library lending navigation system based on collaborative filtering and Augmented Reality (AR) technology is designed. A joint collaborative filtering model based on attention mechanism is also proposed to solve the data sparsity problem of library recommendation system. In addition, the study introduces AR technology for book location marking and circulation navigation in order to provide better circulation service experience in library scenarios. The AR plane detection and ray collision detection technologies enable users to locate the exact position of books through their smartphones. Experiments on Movielens and BookCrossing datasets show that the model exhibits excellent recommendation in both top-K Hit Ratio (HR) and Normalized Discounted Cumulative Gain (NDCG) metrics. The results show that the accuracy localization error is higher than that of the normalized discounted cumulative gain (NDCG), with the improvement ranging from 2.6% to 26.4%, respectively. The experimental results show that the maximum accuracy positioning error is only 0.62 m, and in most cases, the positioning accuracy of the fine positioning is higher than that of the coarse positioning. In summary, the joint collaborative filtering model based on attention mechanism and AR navigation technology can effectively improve the efficiency of library lending service and user satisfaction.