Research on knowledge-driven analytics in intelligent manufacturing

This article provides an integrative review of knowledge-driven analytics in intelligent manufacturing. To address the lack of a unified conceptual framework, it proposes a five-layer model linking data infrastructure, analytics, knowledge representation and reasoning, manufacturing applications, and management outcomes. Methodologically, the study applies bibliometric mapping to 1,542 publications from 2007 to 2025, using CiteSpace 6.3.R4 to analyze keyword co-occurrence, temporal evolution, and citation bursts, complemented by a structured narrative synthesis. The review explains how data acquisition, preprocessing, feature engineering, and complex event processing transform raw industrial data into contextualized information and actionable knowledge. Knowledge-driven analytics creates value when analytical outputs are converted into reusable manufacturing knowledge and embedded in decisions involving quality control, equipment health management, process optimization, supply-chain coordination, and sustainability management. Key challenges include data quality, dynamic knowledge updating, domain knowledge–model integration, and analytics governance. Emerging directions include causal reasoning, neural-symbolic learning, digital-thread knowledge flows, and human-machine collaborative knowledge systems. Overall, its effectiveness depends not only on analytical performance but also on governance, organizational learning, workforce capability, and human-machine knowledge integration.