On the Use of Self-Organizing Map for Text Clustering in Engineering Change Process Analysis: A Case Study.

Area of Science:

• Engineering Management
• Data Science
• Knowledge Management

Background:

• Complex product development necessitates managing engineering changes.
• Engineering Change Requests (ECRs) document proposed modifications but are often underutilized.
• Existing ECR data represents a missed opportunity for organizational learning and knowledge exploitation.

Purpose of the Study:

• To explore the application of Self-Organizing Maps (SOM) for unsupervised clustering of ECR texts.
• To assess the potential of SOM for improving knowledge reuse within the engineering change process.
• To demonstrate the efficacy of SOM text clustering through a case study in the railways industry.

Main Methods:

• Utilized Self-Organizing Map (SOM) algorithm for unsupervised text clustering.
• Applied SOM to a dataset of ECRs from the railways industry.
• Analyzed clustered ECRs to identify patterns and facilitate knowledge extraction.

Main Results:

• SOM effectively clustered unstructured ECR text data.
• Identified recurring issues and best practices within ECRs through clustering.
• Demonstrated the potential for improved knowledge reuse and exploitation.

Conclusions:

• Self-Organizing Maps offer a viable approach for analyzing and organizing ECR data.
• SOM-based text clustering can significantly enhance knowledge management in engineering change processes.
• This method facilitates learning from past projects, leading to more efficient product development.