Introduction. Knowledge graphs (KGs), a method of knowledge organization based on semantic network theory, integrate vast, dispersed knowledge into machine-readable, structured semantic networks. KGs have been systematically reviewed and applied in fields like biomedicine, education, and psychiatry. However, within psychology, despite growing research, a systematic review of KG application models and prospects is lacking.
Crucially, applying psychological knowledge faces a fundamental challenge: its concepts and theories are highly abstract and subjective, lacking formal expression and computable semantic structures. While traditional empirical research reveals statistical correlations, it struggles to systematize these links into a knowledge system with clear semantic relations and logical hierarchies. KGs may offer a new approach for building computable and inferable psychological knowledge structures.
Methods. Based on a systematic literature review and analysis integrating computer science, information science, and psychology (searching databases like PubMed, IEEE Xplore, ScienceDirect, and Google Scholar), this paper synthesizes research to distill key potential application directions for KGs in psychology.
Results. This paper systematically expounds on KG construction methods and potential application pathways, emphasizing the value of key technologies like KG embedding (KGE), knowledge base question answering (KBQA), and node clustering. KGE represents entities and relations as low-dimensional vectors, enabling computers to "understand" psychological knowledge and support predictive modeling. KBQA enhances interactive flexibility by retrieving and organizing information. Node clustering provides a path for deep cognitive mining by quantifying semantic relevance and topological proximity.
Furthermore, integrating KGs with emerging technologies like large language models (LLMs) and digital twins promises to enhance the intelligence and application value of psychological research. The paper also addresses ethical challenges, such as data privacy and algorithmic fairness.
Conclusions. KGs provide psychology with a computable, inferable knowledge infrastructure. Future research should focus on building high-quality psychological KGs, promoting real-world applications, and strengthening ethical standards for technological practice.