Knowledge graph-based intelligent data management and information innovation service model for university library systems

TL;DR

This paper introduces a smart data management model for university libraries using knowledge graphs to improve personalized information services and user experience.

Contribution

A novel hierarchical model integrating knowledge graphs and large language models to enhance library data management and personalized services.

Findings

User experience score improved by 45.2% over six months.

Search accuracy increased by 41.1%, and service quality improved by 32.5%.

All 16 technical indexes met or exceeded set standards.

Abstract

With the rapid development of digitalization, university libraries find themselves under great pressure in adapting to data management, whereas traditional management has limitations in meeting personalized needs for information. This study constructs a knowledge graph-based intelligent data management and information innovation service model for university library systems, which adopts a hierarchical design philosophy encompassing five core layers: data source layer, data processing layer, knowledge construction layer, service application layer, and user interaction layer. By integrating multi-source heterogeneous data resources and establishing a unified knowledge representation framework, the model facilitates semantic organization as well as automatic management of library information. The model employs dynamic fusion methods combining large language models and graph embedding to address heterogeneous data integration challenges, while leveraging knowledge graph semantic association capabilities to provide precise personalized information recommendation services. A systematic evaluation conducted for a period of six months shows that score for an user experience is 4.40, pointing to an improvement of 45.2% from 3.03, with accuracy in search results increasing by 41.1%, as well as enhancement of service quality and learning effectiveness by 32.5% and 41.7% respectively, and all 16 technical indexes having met and exceeded set standards. This study proposes a realistic solution to help university libraries deal with challenges brought by big data, which also facilitates intelligent service transformation for university libraries.

Figures