Extracting and Predicting of Industrial Quality Knowledge with Relation Attention Networks
Actual industrial production is often accompanied by a large number of miscellaneous production data, but these data can not be fully used. This paper provides an idea for mining the potential value of data from industrial production by using data association analysis and knowledge embedding model. Firstly, to get the correlation degree between features from the production data, the grey correlation analysis method is adopted. Then, the product quality correlation degree knowledge graph is formed based on heterogeneous relation attention networks. Furthermore, the unknown quality knowledge is derived by using the knowledge embedding model. Finally, the validity of the applied method is verified with the specified data set.
Authors
- Xinyu Yang (ORCID: https://orcid.org/0000-0003-0481-4880)
- Jinyan Liao
- Dawei Zhang (ORCID: https://orcid.org/0000-0003-4639-8504)
- Yong Chen (ORCID: https://orcid.org/0000-0003-4580-8655)
- Xiaoying Zhang (ORCID: https://orcid.org/0000-0002-0055-7322)
- Longjie Zhang (ORCID: https://orcid.org/0000-0003-3052-5273)
Institutions
- University of Electronic Science and Technology of China (CN)
- Gezhouba Explosive (China) (CN)
Publication Details
- Published
- 2024-11-01
- DOI
- https://doi.org/10.1109/cac63892.2024.10865576
- Primary Topic
- Industrial Vision Systems and Defect Detection
- Type
- article
- Field-Weighted Citation Impact
- 0.00
Funders
- National Natural Science Foundation of China