Article: Personalised book recommendation model for university libraries based on multi-factor knowledge tracking Journal: International Journal of Information and Communication Technology (IJICT) 2025 Vol.26 No.50 pp.1 - 16 Abstract: University libraries are confronted with the challenges of low resource utilisation rate and insufficient modelling of the dynamic evolution of readers' cognition. Traditional collaborative filtering methods are difficult to quantify cognitive state changes and ignore the influence of environmental factors on resource adaptability. To this end, this study proposes a dynamic recommendation model that integrates multi-factor knowledge tracing (MFKT) and graph neural networks (GNN). The reader cognitive state matrix is constructed through gated recurrent unit (GRU) time series modelling. Combined with behavioural pattern analysis and environmental feedback mechanism, the dynamic balance of resource difficulty and popularity is achieved. The cognitive graph convolutional network (CGCN) is designed based on the Pareto optimality theory to synchronously optimise the recommendation accuracy, knowledge gain and resource coverage. This study provides a referable technical solution to solve the problem of accurate matching between resources and readers' cognition. Inderscience Publishers - linking academia, business and industry through research

Title: Personalised book recommendation model for university libraries based on multi-factor knowledge tracking

Authors: Fanglin Deng

Addresses: Library, Guangzhou Maritime University, Guangzhou 510725, China

Abstract: University libraries are confronted with the challenges of low resource utilisation rate and insufficient modelling of the dynamic evolution of readers' cognition. Traditional collaborative filtering methods are difficult to quantify cognitive state changes and ignore the influence of environmental factors on resource adaptability. To this end, this study proposes a dynamic recommendation model that integrates multi-factor knowledge tracing (MFKT) and graph neural networks (GNN). The reader cognitive state matrix is constructed through gated recurrent unit (GRU) time series modelling. Combined with behavioural pattern analysis and environmental feedback mechanism, the dynamic balance of resource difficulty and popularity is achieved. The cognitive graph convolutional network (CGCN) is designed based on the Pareto optimality theory to synchronously optimise the recommendation accuracy, knowledge gain and resource coverage. This study provides a referable technical solution to solve the problem of accurate matching between resources and readers' cognition.

Keywords: multifactor knowledge tracking; MFKT; book recommendation model; graph neural networks; GNN; gated recurrent unit; GRU.

DOI: 10.1504/IJICT.2025.151070

International Journal of Information and Communication Technology, 2025 Vol.26 No.50, pp.1 - 16

Received: 05 Jul 2025
Accepted: 15 Aug 2025
Published online: 12 Jan 2026 *

Title: Personalised book recommendation model for university libraries based on multi-factor knowledge tracking

Keep up-to-date