Research Topic Identification of Supply Chains Empowered by Artificial Intelligence Technologies: Based on LDA Text Mining

Existing research on AI-empowered supply chains remains constrained by fragmented topic clustering, unclear value transmission mechanisms, and the absence of a holistic theoretical framework. Drawing on 4,635 English-language articles published between 2000 and 2026 in ABS 3-star and above journals indexed in the Web of Science Core Collection, this study applies LDA topic modeling to identify research topics and deconstruct empowerment dimensions. Following text preprocessing, the optimal number of topics (K = 20) is determined using perplexity metrics. Twenty core topics are identified and clustered into five empowerment dimensions: digital-intelligence optimization, chain-link operation, supply chain network (SCN) security, cross-border scheduling, and sustainable development. Social network analysis is then used to quantify topic correlations and value transmission paths, based on which a conceptual model of AI-empowered supply chains is developed. The findings show that AI empowers supply chains through these five clearly bounded dimensions, forming a closed-loop value-generation and dynamic feedback mechanism that links theoretical research with industrial practice. By integrating fragmented findings, this study deconstructs the field’s core topics and empowerment dimensions and establishes a closed-loop interpretive framework. It clarifies how AI is transformed from a technical application into strategic supply chain value, offering theoretical references and practical implications for intelligent supply chain transformation.