Scholarlink

Quality 4.0 integrates Industry 4.0 technologies into quality management by analyzing machine sensor data with intelligent systems to predict and identify failures and quality issues. This paper presents a method for diagnosing the causes of product quality problems using sensor data and quality control outcomes. It employs the Concept Induction framework to derive precise, explanatory concepts that distinguish compliant from non-compliant products. A new evaluation metric, combining instance coverage, classification precision, and hierarchical relevance, is proposed to identify the most informative concepts. The approach is validated using a synthetic Predictive Maintenance dataset simulating a milling machine, linking sensor data with quality outcomes via an ontology. Results show improved concept induction and accurate identification of factors related to defective products.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A Novel Concept Induction Approach for Explainable Quality 4.0

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Léa Charbonnier,
  • Franco Giustozzi,
  • Julien Saunier,
  • Cecilia Zanni-Merk

摘要

Quality 4.0 integrates Industry 4.0 technologies into quality management by analyzing machine sensor data with intelligent systems to predict and identify failures and quality issues. This paper presents a method for diagnosing the causes of product quality problems using sensor data and quality control outcomes. It employs the Concept Induction framework to derive precise, explanatory concepts that distinguish compliant from non-compliant products. A new evaluation metric, combining instance coverage, classification precision, and hierarchical relevance, is proposed to identify the most informative concepts. The approach is validated using a synthetic Predictive Maintenance dataset simulating a milling machine, linking sensor data with quality outcomes via an ontology. Results show improved concept induction and accurate identification of factors related to defective products.