{"title":"Exploring topic modelling for generalising design requirements in complex design","authors":"Cheng Chen, Beshoy Morkos","doi":"10.1080/09544828.2023.2268850","DOIUrl":null,"url":null,"abstract":"AbstractAs the redesign process progresses in product lifecycle management, effectively managing engineering changes becomes increasingly challenging, often leading to catastrophic and costly project failures. In response, the study provides a framework for generalising design requirements documents into topics that engineers can use to understand complex designs. Based on previous work, this study employs and compares four different models, including latent Dirichlet allocation (LDA), the collapsed Gibbs sampling algorithm for the Dirichlet multinomial mixtures model (GSDMM), LDA-BERT, and GSDMM-BERT to determine the appropriate representation of requirements documents. Both heatmaps and UMAPs are used to illustrate the correlation between topics and words. The results indicate that the combined vector representation of topic modelling and the sentence-BERT model outperforms single topic modelling. This combined model leverages the additional knowledge from a pre-trained sentence-BERT model, thereby improving model performance and word distribution in all three industrial projects. Through this proposed framework, engineers can potentially generalise high-quality requirements topics for large requirements documents.KEYWORDS: Requirement managementrequirement topicscomplex designBERTdesign process Disclosure statementNo potential conflict of interest was reported by the author(s).Notes1 https://radimrehurek.com/gensim/models/ldamodel.html2 https://docs.scipy.org/doc/scipy/reference/generated/scipy.spatial.distance.pdist.html","PeriodicalId":50207,"journal":{"name":"Journal of Engineering Design","volume":null,"pages":null},"PeriodicalIF":2.5000,"publicationDate":"2023-10-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Journal of Engineering Design","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1080/09544828.2023.2268850","RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"ENGINEERING, MULTIDISCIPLINARY","Score":null,"Total":0}
引用次数: 0
Abstract
AbstractAs the redesign process progresses in product lifecycle management, effectively managing engineering changes becomes increasingly challenging, often leading to catastrophic and costly project failures. In response, the study provides a framework for generalising design requirements documents into topics that engineers can use to understand complex designs. Based on previous work, this study employs and compares four different models, including latent Dirichlet allocation (LDA), the collapsed Gibbs sampling algorithm for the Dirichlet multinomial mixtures model (GSDMM), LDA-BERT, and GSDMM-BERT to determine the appropriate representation of requirements documents. Both heatmaps and UMAPs are used to illustrate the correlation between topics and words. The results indicate that the combined vector representation of topic modelling and the sentence-BERT model outperforms single topic modelling. This combined model leverages the additional knowledge from a pre-trained sentence-BERT model, thereby improving model performance and word distribution in all three industrial projects. Through this proposed framework, engineers can potentially generalise high-quality requirements topics for large requirements documents.KEYWORDS: Requirement managementrequirement topicscomplex designBERTdesign process Disclosure statementNo potential conflict of interest was reported by the author(s).Notes1 https://radimrehurek.com/gensim/models/ldamodel.html2 https://docs.scipy.org/doc/scipy/reference/generated/scipy.spatial.distance.pdist.html
期刊介绍:
The Journal of Engineering Design is a leading international publication that provides an essential forum for dialogue on important issues across all disciplines and aspects of the design of engineered products and systems. The Journal publishes pioneering, contemporary, best industrial practice as well as authoritative research, studies and review papers on the underlying principles of design, its management, practice, techniques and methodologies, rather than specific domain applications.
We welcome papers that examine the following topics:
Engineering design aesthetics, style and form-
Big data analytics in engineering design-
Collaborative design in engineering-
Engineering concept design-
Creativity and innovation in engineering-
Engineering design architectures-
Design costing in engineering
Design education and pedagogy in engineering-
Engineering design for X, e.g. manufacturability, assembly, environment, sustainability-
Engineering design management-
Design risk and uncertainty in engineering-
Engineering design theory and methodology-
Designing product platforms, modularity and reuse in engineering-
Emotive design, e.g. Kansei engineering-
Ergonomics, styling and the design process-
Evolutionary design activity in engineering (product improvement & refinement)-
Global and distributed engineering design-
Inclusive design and assistive engineering technology-
Engineering industrial design and total design-
Integrated engineering design development-
Knowledge and information management in engineering-
Engineering maintainability, sustainability, safety and standards-
Multi, inter and trans disciplinary engineering design-
New engineering product design and development-
Engineering product introduction process[...]