An adaptive semantic retrieval framework for digital libraries integrating graph neural networks, ontology, and user...

Area of Science:

• Computer Science
• Information Science
• Artificial Intelligence

Background:

• Traditional knowledge organization in digital libraries uses static methods, failing to capture complex relationships or adapt to user needs.
• Existing systems struggle with multidimensional relationships and evolving user information-seeking behaviors.
• A gap exists in integrating formal knowledge structures with dynamic user patterns for adaptive retrieval.

Purpose of the Study:

• To develop a novel adaptive semantic retrieval framework for digital libraries.
• To integrate graph neural networks (GNNs), ontological knowledge, and user behavior analysis.
• To enhance information discovery by dynamically balancing domain semantics and personalized relevance.

Main Methods:

• Constructing an ontology-driven knowledge graph.
• Applying multi-relational GNNs for representation learning.
• Developing a user behavior model and an adaptive retrieval mechanism.
• Integrating formal semantics with empirical usage patterns.

Main Results:

• The adaptive framework achieved 81% precision and 85% recall in experiments.
• Significant performance improvements over conventional retrieval models were demonstrated.
• The system successfully balanced domain semantics with personalized relevance signals.

Conclusions:

• The proposed framework offers a unified computational approach for intelligent information discovery.
• It bridges traditional classification with user-centered design principles for adaptive knowledge organization.
• This approach enhances semantic coherence and responsiveness in digital library retrieval.