Application effect of improved retrieval system based on user understanding on library management system

Abstract

As a key function in library management, optimizing the retrieval system has become a top priority. The current retrieval methods cannot accurately understand the user's retrieval needs and objectives. Therefore, a new semantic retrieval model is constructed based on the bidirectional encoder representations from transformer and the knowledge graph. This model improves the retrieval system from two aspects: pseudo-correlation feedback and semantic word expansion. In addition, the model implements policy control for different document and query lengths to further improve the retrieval efficiency. From the results, the recall and precision of the semantic retrieval model were 90.02% and 95.08%, respectively. After one month of putting the model into library management, over 80% of the participants in the questionnaire survey believed that the retrieval system had high retrieval efficiency and strong understanding for the query content during the retrieval process. The designed model can effectively assist library staff in managing information, further improving management efficiency, and providing a better user experience for student borrowing.