Towards a Framework for Identifying Relevant Information in regard to Specific Context on the Use Case of Standards and Directives
DS 125: Proceedings of the 34th Symposium Design for X (DFX2023)
Year: 2023
Editor: Dieter Krause, Kristin Paetzold-Byhain, Sandro Wartzack
Author: Max Layer (1), Sebastian Neubert (1), Brij Boda (1), Ralph Stelzer (2)
Series: DfX
Institution: 1: Siemens Energy Global GmbH & Co.KG; 2: Virtual Product Development, Technical University Dresden
Page(s): 235-244
DOI number: 10.35199/dfx2023.24
Abstract
In a complex landscape of engineering requirements and knowledge represented by standards and directives, navigating and interpreting context-specific information remains a substantial challenge. A novel approach to address this issue is presented, introducing an information extraction framework for identifying product and context-specific, relevant information within unstructured text. The methodology employs Natural Language Processing techniques in a pipeline to parse text from various resources. Different stages for introducing context are proposed based on a trade-off between speed, accuracy, and storage capacity. An initial test focuses on the identification of inspection requirements of piping, while illustrating other potential applications such as an external reference cluster.
Keywords: Text analysis, Standards, Information Retrieval, Clustering, Natural Language Processing