Enriching Analytics Models with Domain Knowledge for Smart Manufacturing Data Analysis

Area of Science:

• Manufacturing Engineering
• Data Science
• Artificial Intelligence

Background:

• Data analytics is crucial for decision-making in Smart Manufacturing.
• Integrating domain knowledge into analytics models is challenging, leading to interoperability and traceability issues.
• Current practices often leave domain knowledge undocumented or poorly integrated.

Purpose of the Study:

• To propose a methodology for enriching analytics models with domain knowledge.
• To address the limitations of current data analytics projects in Smart Manufacturing.
• To enhance the integration and utilization of domain expertise in model development.

Main Methods:

• Development of a novel methodology to enrich analytics models with domain knowledge.
• Implementation of a case study to demonstrate the methodology's application.
• Utilizing a Bayesian Network model for illustrating the enriched analytics approach.

Main Results:

• The proposed methodology successfully enriches analytics models with domain knowledge.
• The case study demonstrated the practical application and benefits of the enriched model.
• Significant improvements in the efficiency of developing Bayesian Network models were observed.

Conclusions:

• Enriching analytics models with domain knowledge is feasible and beneficial.
• The methodology enhances interoperability and traceability of domain knowledge in analytics.
• This approach improves the efficiency and effectiveness of data analytics in Smart Manufacturing.