Advanced Engineering Informatics 54 (2022) 101774

Contents lists available at ScienceDirect

Advanced Engineering Informatics

journal homepage: www.elsevier.com/locate/aei

Full length article

Data-driven engineering design: A systematic review using

scientometric approach
Daria Vlah a, *, Andrej Kastrin b, Janez Povh a, Nikola Vukašinović a
a
Faculty of Mechanical Engineering, University of Ljubljana, Ljubljana, Slovenia
b
Faculty of Medicine, University of Ljubljana, Ljubljana, Slovenia

A R T I C L E I N F O A B S T R A C T

Keywords: In the last two decades, data regarding engineering design and product development has increased rapidly. Big
Data-driven design data exploration and mining offer numerous opportunities for engineering design; however, owing to the
Engineering design multitude of data sources and formats coupled with the high complexity of the design process, these techniques
Product development
are yet to be utilised to the best of their full potential. In this study, a comprehensive assessment of the state-of-
Big data analytics
the-art data-driven engineering design (DDED) in the last 20 years was conducted. A scientometric approach was
Bibliographic analysis
Network analysis employed wherein first, a systematic article acquisition procedure was performed, where a dataset of 3339 ar­
ticles related to engineering design and big data analytics applications were extracted from Web of Science (WoS)
and Scopus. Thereafter, this dataset was reduced to a dataset of 366 articles based on concise data screening. The
resulting articles were used to analyse the dynamics of research in DDED throughout the last 20 years and to
detect the primary research topics related to DDED, the most influential authors, and the papers with the highest
impact in the DDED domain. Furthermore, the co-occurrence network of keywords/keyphrases and co-
authorship networks were constructed and analysed to reveal the interconnection of the research topics and
the collaboration between the most prolific authors. Finally, an insight how big data analytics is being applied
through product development activities to support decision-making in engineering design was presented.

1. Introduction the conditions and constraints of a particular task. It is among the fields
characterised by long traditions and well-established processes. With the
1.1. Motivation rapid technological advancements there is an increased need to reduce
product lead times and expenses, adjust to quickly changing market
Currently, we are in the process of intensive digitalisation, and demands, and to improve the efficiency of processes. Furthermore, the
consequently, the data generated by the millions of items that surround digital era has resulted in design engineers and manufacturers encoun­
us is constantly increasing. The need to assess and exploit these data has tering new challenges:
resulted in the rapid development of methods and technologies known
as big data analytics (BDA). BDA is an umbrella term for science and • The complexity of products has increased as they are no longer
technologies that enable organising big data, analysing and discovering purely mechanical; very often products exhibit a higher level of
knowledge, patterns, and intelligence from big data, and visualising and interaction with the users, which require different design approaches
reporting the discovered knowledge to assist decision-making [1,2]. The owing to their expanded functionalities [6,7].
development of data-enabled technologies coupled with the democrat­ • The utilisation of the new data-based technologies must consider
isation of data science, has enabled the implementation of big data ap­ various characteristics of engineering design work, e.g., the ill-
proaches within different engineering domains, as well as in engineering structured design problems, exploratory and pragmatic nature of
design [3,4]. design processes [8], reliance on judgment and experience [9], and
Pahl et al. [5] defined engineering design as the activity of deter­ the high degree of collaboration between designers [10].
mining solutions for technical problems via application of insights based • Design engineers must leverage vast amounts of product data from
on natural and engineering sciences, while simultaneously considering various sources in a variety of formats. Consequently, big data will

* Corresponding author at: University of Ljubljana, Faculty of Mechanical engineering, LECAD laboratory, Askerceva 6, 1000 Ljubljana, Slovenia.
E-mail address: [email protected] (D. Vlah).

https://doi.org/10.1016/j.aei.2022.101774
Received 24 November 2021; Received in revised form 17 August 2022; Accepted 1 October 2022
Available online 19 October 2022
1474-0346/© 2022 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
D. Vlah et al. Advanced Engineering Informatics 54 (2022) 101774

require new systems engineering and architectural solutions owing Table 1

to the complexity of analysing large datasets [11]. Research questions.
Focus area Research questions (RQ)
Knowledge discovery from data is not a new concept in mechanical
RQ1 Yearly evolution of DDED How has the number of publications on data-
engineering [12]. In the past decades, substantial research effort has research driven engineering design evolved between
been invested into developing methodologies and models for the 2001 and 2021?
collection, organisation, and representation of design data in RQ2 Main topics in DDED What are the primary topics associated with
knowledge-based systems that could capture the design intent, for big data analytics applications in engineering
design?
example, [13–16]. Such systems containing design information can RQ3 Authors, productivity, and How are the authors of DDED publications
enable the manipulation and reuse of the stored design knowledge, impact distributed geographically, how do they
aiding in decreasing the required effort and time to design [17]. collaborate, and what are the most frequently
Scientific publications highlighting BDA applications in engineering cited publications related to DDED?
RQ4 Big data analytics In which phases of the product development
design have begun to increasingly emerge only recently owing to the
applications in product processes can we find examples of big data
complexity of the design process, dispersed and diverse sources of design development analytics applications? What are the main
data, and the availability of simple tools and methodologies that can activities in the design process that can be
allow usage of BDA without specialist data science expertise [18]. Data- enhanced with the use of big data analytics?
driven engineering design (DDED) is a new paradigm in engineering What data sources can be used?

design where the data are extensively utilised to support decision-

making [19,20]. It uses BDA techniques to discover relations and hid­ authorship network, identified the most productive communities,
den patterns from product and customer data, and to extract valuable and created a list of the 20 most cited publications in DDED.
information from sources such as e-commerce webpages [19,21–23], • Insights how BDA is applied in product development activities to
social media [24], product and lifecycle management systems (PLM) support engineering design decision-making.
[25], 3D CAD and simulation models [26,27] and sensors [28]. The
potential of DDED has been recognised by engineering practitioners
[29,30]; however, as research on DDED is still in its nascent stages, there 1.3. Related work
is a need to synthesise the existing knowledge and to provide an over­
view of the different capabilities of DDED and thus also encourage In the last few years, various aspects of data-driven approaches in
further research in DDED and its implementation in industry. engineering design have been researched upon. However, currently,
The motivation for the study was to investigate the advantages of there exists no commonly accepted definition of data-driven engineering
using data from various sources and BDA techniques in various aspects design (DDED) because of the novelty of approaches integrating data in
of engineering design, and to provide an overview of the research con­ engineering. Consequently, several different terms and frameworks
ducted up to date and current trends. A systematic review was employed considering various aspects of the design and product development
using the scientometric approach. This approach was chosen because it process have been proposed in the literature [32]. Van Horn et al. [33]
enables structured and data-based exploration of the state-of-the-art. It introduced the term design analytics as a new paradigm for improving the
has not yet been widely adopted in reviewing results of engineering transformation of design information into knowledge based on extensive
(sub)domains; hence, explicit guidelines, methodologies, or automation use of user and design-related big data. Liu and Chen [19] defined data-
tools that would support this process for engineering design research are driven design as a decision-making framework wherein the involvement
scarce [31]. Thus, this study also attempts demonstrate the potential of of data in the design process can support achieving optimal decisions.
systematic review approach to present the state-of-the-art in data-driven Feng et al. [20] described data-driven product design as the process of
engineering design. mining the relevance and hidden patterns of things via modelling and
analysis of product-related data to assist product design. Based on the
product design process, they developed a research framework incorpo­
1.2. Research questions and contribution rating customer requirement analysis, conceptual design, detailed
design, and design knowledge support tools. Chiarello et al. [34]
This study investigates by systematic review (i) how the DDED explored the intersections between data science and engineering design
research has developed through the years, and (ii) what are the benefits considering a scoping review supported by natural language processing
of using BDA in engineering design, through four research questions. techniques. A variety of intersecting areas were discovered, including
They cover the yearly dynamics of the related research, key and the primary stages of the design process (problem definition, concep­
emerging topics in DDED, leading authors and their main contributions, tual, embodiment, and detail design), design education, design for X,
and the way BDA techniques are applied in different product develop­ innovation and IP management, project management, prototyping, and
ment activities. These questions are formally stated in Table 1. sustainability. First, Chiarello et al. limited their study to Scopus only,
The main contributions of the presented article are: whereas we also included Web-of-Science data. Second, they used the
word embedding approach to expand the search space for relevant ar­
• Systematic article acquisition procedure, wherein from Web of Sci­ ticles, while the comparative advantage of our approach is the manual
ence (WoS) and Scopus, a dataset of 3339 articles related to engi­ review of the keyword list. Bertoni [35] conducted a systematic review
neering design and big data analytics applications were extracted. of data-driven design, focusing on the conceptual design phase. The
Following a precise data screening, this dataset was reduced to 366 analysis showed that most studies on concept development have focused
articles clearly related to DDED. on customer need identification, while other activities such as product
• Computed and visualised data on the evolution of research in DDED concept generation, design space exploration, and establishing specifi­
over the last 20 years, measured by the number of publications. We cations have not been explored much. Text mining was found to be the
identified the most important topics in DDED by analysing key­ most prevalent data extraction method, and internet sources, including
words/keyphrases (henceforth referred as keywords). The network online reviews and social media, were the most common data sources.
of keywords that we created indicates how the topics were Shabestari et al. [36] provided a brief overview of the state of the ap­
interconnected. plications of machine learning in the early product development phases,
• Identification of the most productive countries in terms of co- considering requirement analysis, functional modelling, and principal
authorship of publications on DDED. In addition, we created a co- concept design.

2
D. Vlah et al. Advanced Engineering Informatics 54 (2022) 101774

Product development and product use cannot be distinguished [37] based overview of the state-of-the-art data-driven engineering design.
and it is argued that the use phase should be included as part of the
process [33]. The products in use can provide designers with valuable 2. Methodology
information regarding the different usage scenarios, unidentified or
misinterpreted user needs, product quality, and defects. Recently, Owing to the lack of formalised instructions, to conduct the sys­
several approaches and frameworks have been proposed by researchers tematic review, we adopted approaches from other domains such as
that investigate data-driven engineering design, focusing particularly on healthcare and software engineering [45,46] and adjusted them ac­
the use phase [38]. According to [39] the term data-driven design tends cording to our research needs based on the guidelines suggested by Lame
to initiate discussions regarding the use of data to optimise and refine [47] and Torres-Carrion et al. [48]. The study was conducted using a
products in relation to easily quantified performance metrics. The use of quantitative, bibliometric, and scientometric approach for systematic
sensors on prototypes and connected products has been suggested as review synthesis. The methodological framework of the study was
potential sources of data for describing the user-product interactions. organised according to a general data analysis procedure [49] and
Further, Zheng et al. [40] used the term data-driven product development consisted of five steps (Fig. 1). To answer the research questions, first,
from the perspective of the industrial Internet of Things, referring to we collected all available articles from the selected scientific databases
methods using which robust designs that respond to customer demands automatically, and then carefully selected relevant articles to perform
can be achieved by utilising product data. The suggested approach aims the data analysis to conduct this systematic literature review. The pro­
to help designers enhance their organisations’ competitive edge by cess of the collection of articles is described in Section 2.1. Screening,
uncovering business patterns, novel insights, and implicit knowledge. that is, the selection of relevant articles on the topic of DDED, is
Finally, Trauer et al. [32]. suggested a data-driven engineering framework described in Section 2.2. Methods for the retrieval, information extrac­
for technical product development wherein the use-case-oriented tion, analysis, and visualisation of the data obtained from the collected
collection and utilisation of sufficiently connected product lifecycle articles are described in section 2.3. Further, for each of the four posed
data guides as well as drives decisions and applications in the product research questions, different data from the retrieved articles were
development process. Wang et al. [41] reviewed the theories, algo­ extracted, and several analysis and visualisation methods were applied
rithms, and technologies of AI for various stages of product lifecycle for a clearer representation of the outputs.
management (PLM): product design, manufacturing, and service. They
reported that in the product design stage, AI can enhance design 2.1. Article acquisition procedure
decision-making in conceptual, embodiment, and detail design by
mapping the information extracted regarding customer needs and A literature search and screening strategy was designed based on the
preferences for product attributes, functions, and performance. review protocol proposed by Ali et al. [50]. The research questions,
Certain authors have presented more generalised approaches to defined in Section 1.2, indicated the research scope and construction of
DDED, which include several aspects of stakeholder and data in­ the search query for data retrieval from large article databases. The
teractions. Li et al. [42] proposed an engineering design process model paper follows the standard scientometric analysis procedure [51]. We
that incorporates a data analytics process model into the new product used Web of Science (WoS) and Scopus as the primary sources of the
development process. The model considers key tasks, data sources, data/ publications included in the review, because they contain a wide variety
information flows, and interactions between engineers and data scien­ of papers from many publishers and enable the retrieval of various in­
tists during product design. Cantamessa et al. [3] investigated how formation regarding the articles. The result of the search query also
digitalisation affects designers, design processes, and the role of data- contained articles that did not match our research interest, thus, we
enabling technologies at the individual, organizational, and manage­ defined certain inclusion and exclusion criteria to improve its precision.
rial levels. The proposed conceptual framework describes the relation­ Early works connecting BDA and engineering design were published
ships between objects and actors in the design process, which should be early in the 2000s, for example, [52,53]; hence, in this study, we
considered when developing new supporting tools and methods in included publications published in 2000 or later.
agreement with the data-driven engineering design paradigm. Moreover, to identify a comprehensive and unbiased set of potential
Most of the contributions investigating DDED have been published in keywords, we applied the semi-automatic approach recently described
the past three years, indicating the attention it has garnered in engi­ by Grames et al. [54]. First, for a naive search of relevant bibliographic
neering design. Although there are several contributions that aimed to records, we used a set of expert-defined terms with the following query:
summarise the state-of-the-art in this field, to the best of our knowledge, (‘engineering design’ OR ‘design process’ OR ‘product design’) AND.
none of the reviewed papers provided an overview of the research (‘data-driven design’ OR ‘data-driven’ OR ‘data mining’ OR ‘big data’
development through the years or investigated the leading researchers. OR ‘data analysis’ OR ‘machine learning’).
In addition, the current studies have focused either on the early design The naïve search strategy yielded 1310 records. Subsequently, we
phases [35,36], the four primary phases of the design process [26,42], or used the RAKE algorithm [55] to expand a set of relevant keyword terms
the product lifecycle [4,41]. Product data are generated throughout the considering the title, abstract, and keyword fields of the retrieved
entire product lifecycle, which, along with product development, in­ bibliographic records. Thereafter, the list of 507 terms was manually
cludes product distribution, storage, use, maintenance, and disposal inspected and filtered followed by creation of queries for the WoS and
[43]. Design is considered a key and critical stage of product lifecycle Scopus search engines. The search with this query yielded 3339 records
[44], and there are many places within the design lifecycle where data including journal and conference articles and a few book sections. As the
can contribute to the process, from the early conceptual and re­ words ‘design’ and ‘engineering’ are used in various areas such as drug
quirements gathering stages through optimisation and maintenance design, medicine, construction, education, energy engineering and
[39]. While papers examining the entire product lifecycle entail insights many others, several extracted articles did not correspond to our
into product data that can be used in design, identified studies involving research topic. Therefore, the search results were filtered to discard
this approach have focused on other aspects, such as, smart irrelevant articles. Two inclusion criteria were applied: (1) contributions
manufacturing [4,41] or product-service systems [7], but lack more that are relevant to the engineering design domain, and (2) contribu­
detailed information on data utilisation during design activities. Thus, tions must include BDA, data mining, or machine learning applications
there is a requirement for a comprehensive overview regarding the state in engineering design or should include reviews of DDED approaches.
of research through the product development (PD) framework, based on This was implemented by limiting the search to publication sources that
which the applications of data mining methods can be studied [36]. In targeted the engineering design domain. First, a list of renowned jour­
this study, we aim to address these concerns and provide an evidence- nals and conference proceedings for that domain were created;

3
D. Vlah et al. Advanced Engineering Informatics 54 (2022) 101774

Fig. 1. Methodological framework.

thereafter, this list was expanded with additional sources that were consisting of 3339 articles, resulting in 445 articles.
found through the initial search. The final list of sources is presented in
Table 2. Source filters were applied to the initial constructed database
2.2. Screening of the articles

Table 2 The screening procedure applied in this study was conducted ac­
List of journal and conference proceedings that were included in the literature cording to the PRISMA guidelines for systematic reviews [56] and was
sources. Number of obtained articles from each source is stated in the second and presented using a modified template reported by Page et al. [57] (Fig. 2).
the fourth columns. Following the article acquisition, the procedure for the assessment of the
Journal No. Journal No. output data quality was defined. For the presented case, this corresponds
Journal of Computing and 21 International Journal of Computer 7 to the assessment of the relevance of the extracted articles for the
Information Science in Integrated Manufacturing research topic. After filtering the articles by source, 445 articles were
Engineering included in the database. However, before beginning the quantitative
International Journal of 18 Applied Sciences 5
data analysis, the relevance of the refined search results was checked,
Production Research
International Journal of 18 International Journal of Industrial 5 that is, articles that correspond to data-driven engineering design. First,
Advanced Manufacturing Ergonomics the 20 top articles with the highest number of citations (TC) were
Technology examined. Eight were found to be irrelevant to our research topic (8/20
IEEE Access 17 Artificial Intelligence for 3 = 40 %). Although this is a high percentage, it included a small dataset.
Engineering Design Analysis and
Hence, we decided to examine all the articles that had ≥ 10 citations to
Manufacturing - AI EDAM
Advanced Engineering 15 Design Studies 3 check whether there was still a high percentage of irrelevant articles. In
Informatics case this was so, the search query or filtering method would require
Expert Systems with 15 International Journal on Interactive 3 improvement. There were 111 articles with a TC ≥ 10. After manual
Applications Design and Manufacturing - IJIDeM
screening of the article titles and abstracts where necessary, of the 111
Journal of Mechanical Design, 11 Engineering with Computers 3
Transactions of the ASME
articles, 23 were found to be irrelevant (~20 %). Nevertheless, we
Computer-Aided Design and 10 International Journal of Applied 3 manually excluded irrelevant articles based on the analysis of their ti­
Applications Engineering Research tles, which resulted in a final set of 366 publications for further analysis.
Journal of Engineering Design 9 Ergonomics 2 This set of articles was used to answer RQ1, RQ2, and RQ3. In contrast,
Research in Engineering 8 International Journal of 2
to approach RQ4, the set of 366 articles was further reduced to 249
Design Engineering and Applied Sciences
Engineering Applications of 7 Journal of Design Research 1 articles that corresponded to defined product development activities. A
Artificial Intelligence more detailed description of the data preparation for each research
Applied Ergonomics 7 question is provided in the following section.
Conference proceedings
Proceedings of International 37 IFIP Advances in Information and 10
Conference on Engineering Communication Technology
Design (ICED) 2.3. Data extraction, analysis, and visualization
Proceedings of the ASME 36 Proceedings of NordDesign 5
Design Engineering Certain data necessary for the analysis of the publications were
Technical Conference retrieved directly from the bibliographic records, while others were
Procedia CIRP 26 IEEE International Conference on 5
Industrial Engineering and
extracted through text analysis. Data were prepared according to the
Engineering Management literature review approach proposed by Albers et al. [58]. The following
ACM International Conference 19 CIRP Annals 3 data were automatically retrieved for each article: authors, authors’
Proceeding Series geolocation, publication year, title, authors’ keywords and expanded
Proceedings of International 15 Proceedings of the Human Factors 3
keywords, journal or conference proceedings title, and citation score.
Design Conference and Ergonomics Society
(DESIGN) Furthermore, the authors, two with a background in engineering design
Conference on Human Factors 11 and two with data science, assessed the additional elements of the
in Computing Systems - collected contributions: product development phase activities associ­
Proceedings ated with the articles, application domains, BDA techniques applied, and
Book chapters 4
the data sources used.

4
D. Vlah et al. Advanced Engineering Informatics 54 (2022) 101774

Fig. 2. PRISMA diagram of the article selection process applied in the systematic review.

2.3.1. Data-driven engineering design research overview occurrence network was constructed based on the keywords of the
When the set of included articles was finalised, that is, all the authors that were retrieved from the final set of 366 articles. A total
remaining articles were relevant for our research, we extracted data of 981 keywords were extracted. First, the keywords were normal­
specific to each research question from the articles. The objective of the ised by domain experts (e.g., ‘knowledge extraction’, ‘knowledge
first three research questions was to provide an overview of data-driven elicitation’ and ‘knowledge acquisition’ were normalized to
engineering design research through the analysis of publication emer­ ‘knowledge extraction’). Further, in the keyword co-occurrence
gence, key topics, prominent authors, and contributions. Consequently, network, nodes refer to terms, and edges connect pairs of nodes
a quantitative approach was applied using the article data retrieved that occur in the same publication. In addition, the weight of the
directly from the databases. The following data strategy was edge indicates the number of times the co-occurrence relation was
implemented. observed in the corpus.
3) RQ3: To identify the key authors and contributions for DDED, data
1) RQ1: The annual production of the articles was presented using the on authors, affiliation/country, article titles, and citations were used.
number of articles published each year from 2001 to 2021. Infor­ Based on the 1st authors’ affiliation, the geographical dispersion of
mation on the publication years and the type of publication (journal researchers was presented through map visualisation. Further, to
articles, conference articles, or book chapters) for each instance was identify the most prolific authors and research groups, we con­
used. structed a co-authorship network indicating the authors who
2) RQ2: The main topics in DDED were identified based on the key­ collaborated on the articles. Out of 366 publications, we disambig­
words most frequently proposed by the authors (Author keywords) uated the authors’ names for 324 publications. The final list of au­
and the frequent words and phrases that appeared in the titles of thors included 967 unique names. In the co-authorship network, the
references cited by the authors (Keyword-Plus). Keywords-Plus were nodes represent the authors, and the edge between the nodes exists
retrieved from the WoS database, providing them for research pur­ when the authors have co-authored a publication. In addition, the
poses to facilitate researchers in locating and retrieving more weight of the edge refers to the number of articles co-authored by a
extensive and relevant information. Furthermore, a keyword co- particular pair of authors. The Louvain clustering algorithm was

5
D. Vlah et al. Advanced Engineering Informatics 54 (2022) 101774

applied to the network partitioning. However, for clarity, the pre­ Table 3
sented co-authorship network (Fig. 5) shows only the largest groups Keyword sets related to the activities in the product development process. *
of co-authors, which counts more authors with multiple publications. Includes different grammatical forms.
Consequently, the most prolific authors, with the highest number of PD Activity Descriptive keywords
publications, were placed central to the research groups in the phase
network. Major research groups were further analysed by article ti­ 0 Market research market analysis, market size, trend map,
tles, and their primary research focuses were extracted. Furthermore, technology space map, emergent technologies,
key contributions in DDED were extracted based on article citation market segment, competitive products,
competitive analysis, benchmark, competition,
scores, and the research topics were summarised.
demand, positioning
0 User research user*, customer*, lead user*, need*, preference*,
2.3.2. BDA applications in product development customer profile*, customization, kansei
To approach RQ4, the retrieved articles were correlated with the engineering, kansei method, affective design,
defined elements of product development. The keyword-based search voice-of-the-customer, VOC, customer
expectation*, ecommerce, online review*,
method [59] was applied using an advanced filtering option on the customer review*, consumer*
initial database. First, the authors consulted the literature on engineer­ 0 Definition of specification*, requirement*, product feature*,
ing design theory to identify specific activities in the product develop­ requirements attribute*, product characteristic*, task
ment process and their characteristics. For each product development definition, design definition, problem
formulation, house of quality, quality function
activity, a list of corresponding keywords was formulated, which was
deployment
further supplemented with findings from previous research questions. 1 Function analysis function*, patent*, TRIZ, functional
Consequently, the keywords were used as inputs for the computer- decomposition, function representation, function
guided categorisation of articles into different product development to behaviour mapping, functional reasoning,
activities according to the article titles. Following this procedure, 249 function-behaviour-structure
1 Principal concept product concept*, concept generation, idea
articles were categorised. Thereafter, the use of different BDA tech­
development generation, creativity, working principle*,
niques in product development activities was explored. BDA techniques problem solution*, concept solution*, solution
were listed from books and papers that provide general overview and design, concept selection, concept evaluation,
categorisation of techniques, and an analogue keyword-based search concept assessment
2 Embodiment design embodiment, layout, product architecture,
was performed to identify articles that contained the applications of
arrangement, preliminary design, design
different BDA techniques. A more detailed description of the process is alternative*, configuration, form, shape, part
provided below. geometry, design exploration, design space,
generative design, topology
2 CAD and CAE CAD, computer aided design, 3D model*, virtual
2.3.2.1. Product development process. According to Ulrich and Eppinger
model*, parametric design, FEM, finite element,
[60], a product development process encompasses a set of activities CAE, computational simulation*, numerical
ranging from market opportunity perception to the production, sale, and simulation*
delivery of a product, which are distributed through different product 2 Product platform product family, product platform, product
development phases from product planning to production. Product assessment portfolio, product variant*
3 Detail design material*, tolerance*, technical documentation,
data are generated and can be utilised in various phases and activities of
drafting, technical drawing*, bill of material,
the product development process (PDP). Considering the inputs from BOM, standard part*, assembly, detail*, MBD,
related work regarding the role of the use phase in data-driven product design for manufacturing
development, this phase was added at the end of the traditional devel­ 4 Testing and field testing, user trial*, reliability, durability,
refinement engineering change*, change propagation,
opment process (Fig. 3).
prototype*, experiment*
In the initial screening of the articles by titles, it was found that the 5 Production production, manufacturing, production cost*,
topics of the articles were more commonly related to specific design- production time, inspection, quality control,
related activities rather than the development phases; for example, machine*, tooling, forming, mould*, equipment,
identifying product requirements, performing functional analysis, and surface, flaws, fault*, defect*, procurement,
supply
conducting engineering simulations, among other things. In addition,
6 Product use use phase, product use, usage, operation,
new agile approaches to product development suggest that activities are sensor*, user behaviour*, digital twin, user
iterated throughout product development, rather than following a linear feedback, service, warranty, user-product
course. Therefore, we decided to explore BDA applications based on interaction*, human–machine interaction*
activities, which can be understood as part of the broader development
phases.
explicit information on design activities.
To explore the use of data-driven approaches for specific activities,
we defined the keyword sets by integrating information from the liter­
2.3.2.2. Big data analytics techniques. Sun and Huo stated that BDA in­
ature describing product development process models and the corre­
cludes mostly data mining (DM), machine learning (ML), data science
sponding design activities coupled with insights from the first three
and systems, and artificial intelligence. In addition, it is also related to
research questions (Table 3). The literature included the PD model by
distributed computing and systems, and cloud computing. The bound­
Ulrich and Eppinger [60], the stage-gate process model [61], and the
aries between BDA, DM, and ML are not clear and are often used
practical engineering guidelines of the German Association of Engineers
interchangeably in engineering design research. Thus, in this study, we
(VDI) [62]. Other product development models do not provide sufficient
used the term BDA techniques for all techniques and algorithms that in

Fig. 3. Product development phases relevant for DDED. Modified model from Ulrich, Eppinger (2019).

6
D. Vlah et al. Advanced Engineering Informatics 54 (2022) 101774

different literature correspond to either big data, data mining, or ma­ We can observe a strong growth of DDED-related publications after
chine learning. Dhall et al. [63] provided a description and classification 2015. A similar trend of growing research in data science and big data
of supervised (labelled data required) and unsupervised (data labelling can also be observed in other areas, such as finance [66], business [67],
not needed) ML techniques and algorithms. The algorithms can be innovation management [68], manufacturing [69], marketing [70], and
grouped according to the types of tasks they can perform into the user experience [71]. The development of tools and methods for the
following categories: regression, classification, clustering, dimension­ acquisition of data regarding products, users and their environment,
ality reduction, and deep learning. Several literature sources were new algorithms, approaches, and tools for big data analysis [72], as well
consulted to construct a list of the most important BDA techniques, as the enhanced computing power and improved storage capabilities
including research papers [2,63] and books [64,65]. enabled easier integration of data science in various fields, thus sup­
The most common algorithms and techniques whose applications porting the rapid growth of data-driven technologies. In addition, in
were found in engineering design include linear and non-linear regres­ engineering design, improved product data management tools and
sion, decision trees (DT), random forest, support vector machines platforms have influenced the faster adoption of such technologies. With
(SVM), naïve Bayes algorithm, k-nearest neighbours (K-NN), artificial the new tools, the processes and designers’ actions can be easily traced
neural networks (ANN), k-means clustering, density-based clustering, along with the decisions made during product development and
hierarchical clustering, fuzzy logic, association rules, principal compo­ manufacturing operations. Furthermore, real-time data acquisition
nent analysis (PCA), factor analysis (FA), convolutional neural network technologies and greater storage capacities can facilitate the monitoring
(CNN), reinforced neural network (RNN), and restricted Boltzmann of products during usage and record user-product interactions. More­
machine (RBM). Two additional techniques considered were: text min­ over, various web harvesting techniques have expanded the possibilities
ing and natural language processing (NLP), because of their significant for the allocation and retrieval of diverse product and user data.
role in early design phases, where the product requirements and func­ Consequently, numerous opportunities for further research are possible;
tions were not yet described quantitatively, and thus, the design- hence, the number of contributions in this field is expected to expand
relevant data must be extracted from unstructured textual sources further in the following years. The yearly increase in the number of
such as engineering documents or internet webpages. research articles on data-driven engineering design appears to be
Using a database comprising 249 articles categorised according to aligned with the growing trend of AI development, as shown by Wang
activities, the titles and abstracts were searched based on keywords, et al. [41].
which highlighted a specific BDA technique. For example, when
searching for articles applying the decision tree technique, the keywords 3.2. RQ2: Main topics in DDED
‘decision tree’, ‘DT’, ‘ID3′ , ‘C4.5′ and ‘tree’ were included. Subsequently,
we reviewed the outputs and further divided the articles, if necessary, Data-driven engineering design is a vast research area that covers
for example, decision trees could be sorted into DT for regression or DT diverse topics. In this study, the main topics occurring in DDED were
for classification. An analogue procedure was applied to all BDA investigated using the keywords extracted from the collected articles.
techniques. The results show frequent terms and common application domains of
BDA in engineering design, which reveal the largest, most researched
3. Results and discussion areas and topics. The performed analysis presents the knowledge map­
ping of existing research in DDED, however, the identification of
3.1. RQ1: Yearly evolution of DDED research emerging trends would require additional analysis.
Fig. 5 shows the co-occurrence network of authors’ keywords in
With the first research question, we aim to investigate the emergence data-driven engineering design articles. The network visualizes the most
of articles related to DDED in the past 20 years. The publication year and frequently used terms and their relations, indicating which keywords
publication type were extracted for each of the 366 relevant publica­ often occur together and the structure of the research topics in DDED.
tions. Fig. 4 shows the number of publications for each year over the past The vertices correspond to the keywords and the links (connections)
two decades. The literature search was performed on 28th of June 2021 between them represent the articles in which two keywords co-occur.
and includes all relevant journal articles, conference papers, and book Further, the size of a vertex represents the degree of the vertex, that
chapters published up to that date. Owing to the time delay in article is, the number of connections to a particular keyword. The network in
indexing, the number of published articles for the years 2020 and 2021 Fig. 5 includes only the vertices with degrees ≥ 20, to render the rep­
is lower than in the previous two years. In addition, in 2020, there was a resentation clearer. However, the actual network contains many more
decline in the number of conference contributions, probably due to keywords and relations than the condensed visualisation.
decreased interest in scientific conferences held online during COVID-19 Keywords ‘machine learning’, ‘data mining’ and ‘big data’ have the
pandemic. largest number of co-occurrences. Their applications in engineering

Fig. 4. Yearly dynamic of publications related to data-driven engineering design (obtained on 28th of June 2021).

7
D. Vlah et al. Advanced Engineering Informatics 54 (2022) 101774

Fig. 5. Co-occurrence network of keywords assigned by authors in the data-driven engineering design.

design are further described in RQ4. ‘Big data’ and ‘machine learning’ Trauer et al [36] referred to such applications of BDA techniques in
are commonly joined with ‘product design’ and ‘product development’. design as data-driven engineering. Lower part of the figure points out
‘Data-driven design’ often comes together with ‘data mining’, ‘big data’ another topic, which correlates ‘big data’ to ‘manufacturing’, ‘smart
and ‘product-service systems’, while ‘machine learning’ has strong manufacturing’, ‘digital twin’, ‘product lifecycle’ and ‘cyber-physical
connections to ‘artificial intelligence’. One strongly interconnected systems’. This implies to the research on the data-driven engineering
cluster includes keywords ‘customer reviews’, ‘kansei engineering’, design where the focus is on utilizing data from the manufacturing and
‘sentiment analysis’, ‘affective design’ and ‘opinion mining’. These consequent stages of the product development for improved product and
keywords include the most researched data source and several analysis service design. Keywords that are most frequently listed together are
methods and approaches used in early design stages, where user pref­ connected by thicker lines. This provides information on the related
erences are investigated. BDA techniques surrounding that cluster are concepts in DDED. Referring to the most frequent keywords in the Fig. 5,
‘text mining’, ‘association rules’, ‘decision tree’ and ‘natural language we can identify major topics in DDED.
processing’, which are suitable for text analysis. On the left side the
keyword ‘creativity’ is related to ‘design process’, and the two tech­
niques, ‘clustering’ and ‘text mining’, which are further adjacent to 3.3. RQ3: Authors, productivity, and impact
‘conceptual design’ stage. On the upper right side of the figure, there is a
group that connects ‘machine learning’ to keywords ‘computer-aided Within the third research question, we explored the most prolific
design’, ‘CAD’, ‘design optimization’, ‘simulation’ and ‘data-driven en­ authors and papers with the highest citation scores with the aim of
gineering’. These terms can be related to the embodiment design phase, identifying the most influential works in DDED. Using the data on author
where the virtual prototyping takes place using CAD and CAE tools. names, affiliations, number of citations for each article, and article titles,
several analyses were conducted to discover the geographical

8
D. Vlah et al. Advanced Engineering Informatics 54 (2022) 101774

distribution of the researchers active in DDED, the most productive follows:

authors, and the most influential papers. Consequently, the qualitative
analysis of the results revealed the application domains of data-driven • Research group, where Kwong C. K. shows greatest visibility, pre­
approaches in design that attracted the most research attention. dominantly involves researchers from Hong Kong and Singapore,
who investigate the acquisition and mining of customer data to assist
3.3.1. Most prolific authors the affective product design [73–75], whereas publications by re­
Fig. 6 shows the geographical dispersion of publications investi­ searchers surrounding Liu Y. focus on data mining from online
gating data-driven engineering design. It depicts countries with two or sources, for example, mining customer reviews for affective design
more publications. Publications were assigned to a particular country [22,76–78], and mining technical documents for conceptual design
according to the 1st author’s affiliation. We excluded five out of 366 [79].
publications from the analysis, as affiliations could not be unambigu­ • Group led by Tao F. and Liu A., attributed mostly to China, investi­
ously identified. The size of the circle that is placed in each country gated the applications of digital twins and AI in product development
corresponds to the number of articles produced per country. Authors in collaboration with Chen C.H.[41,77,80]. In contrast, the group
from the USA published 80 papers and from China, 76 papers. They were mostly represented by Hong Kong’s researchers, Chen C.H., Zheng P.
followed by Germany (40), the United Kingdom (31), and Singapore and Wang Z. is involved in research exploring data-driven ap­
(21). It is evident that the USA and China dominate in the research and proaches to smart product-service system design (PSS) [81–83].
applications of data-driven engineering design, which is consistent with • A group where Luo J., Song B., and Wood K. show the greatest vis­
the findings of Sun and Huo’s study on the broader BDA research area ibility, gathers majority of researchers originating from Singapore,
[2]. who researched data-driven methods, tools, and expert systems for
Fig. 7 presents the co-authorship network of the most productive innovation, specifically focusing on the exploration of patent data for
authors under the broader field of data-driven engineering design ac­ ideation and conceptual design knowledge discovery [84–87].
cording to our dataset. The vertices of this network indicate the authors. • A group of academics from Guangdong, China, including Wang J.,
Two authors are connected to each other with a line if they have pub­ Tian Z., Li Z., and Wang W.M., focused on mining the affective fea­
lished at least two papers together. The size of the vertices corresponds tures and responses in products from online reviews using various
to the number of connections one author has with the others, while the machine learning techniques [21,23,88,89].
width of the link is proportional to the number of articles two authors • Research group surrounding Wartzack S. and Schleich B. is the most
have together. In this network, the connections with the minimal weight represented for DDED in Europe, focusing on a wide area of digital
of 7 are shown, resulting in 116 authors being connected in 11 clusters. engineering, including design knowledge extraction using semantic
The network is shown in Fig. 7. models [90,91], use of AI in computer-aided design (CAD), detailed
Co-authorship of the articles indicated several outstanding research design [32,92], and data-driven approaches for human-product in­
groups (Fig. 7). The largest and highly interconnected groups comprised teractions [93].
the authors in the East Asian geographical area, with four leading • The USA has the highest number of publications; however, there are
research groups emerging from Singapore, Hong Kong, Beijing, and far fewer collaborations between the authors. Tucker C. and related
Guangdong, China. Other prominent research groups included re­ researchers published the most articles, comprising a range of topics,
searchers from Germany and the United States. In addition, the co- including data-driven sketch generation [94], feature exploration
authorship network shows that currently, researchers in DDED are [24,95], portfolio optimisation [96], prototyping [101], and design
quite dispersed and primarily collaborate within their groups. A brief communication [98].
description of the research topics investigated by the largest groups is as

Fig. 6. Geographical dispersion of researchers exploring DDED.

9
D. Vlah et al. Advanced Engineering Informatics 54 (2022) 101774

Fig. 7. Network of co-authors. Authors with the same colour belong to the same group.

3.3.2. Most cited publications The second significant research topic, occurring in several top-cited
Furthermore, we investigated the most-cited articles in DDED. The DDED papers [76,95,102,106,108,110–112], is consumer opinion
papers with the highest number of citations are listed in Table 4. We mining of products for affective design. New sources of data regarding
analysed the publication years, citation scores, article titles, and ab­ user preferences and related product features are being extensively
stracts. The list of top cited works comprises a variety of articles, ranging explored using different text mining and machine learning techniques
from 2002 to 2019. The overview of the most cited articles reveals for their retrieval, for example, online reviews. Kansei engineering is a
several topics in DDED indicating divergent research directions. common approach for mapping user preferences to product design
Three contributions in the top five citations are from 2018 features.
[99,101,114], all related to smart manufacturing, thereby indicating a One of the earliest applications of BDA in product design is product
significant research direction within data-driven engineering. In the family and portfolio design [53]. Using various BDA techniques, product
context of Industry 4.0, the application of sensors paired with IoT, and variants can be compared based on their similarity and evaluated ac­
cloud-based systems have enabled easier collection, storage, and uti­ cording to the way they respond to technical, economic, and user re­
lisation of various manufacturing data, allowing for improved moni­ quirements. Furthermore, data-driven approaches can also be used in
toring and optimisation of the machines and manufacturing processes the analysis of product functions and as a support for developing design
[100] and control [109]. In recent years, the concept of digital twins has concepts [113].
substantially impacted manufacturing. It refers to the creation of a vir­ Two additional identified research directions from the top cited ar­
tual twin of a product or a system that is fed by real-time data, with the ticles are the exploration of ontologies and semantic networks for
aim of obtaining a more realistic image of the system behaviour. capturing the design knowledge from existing design data [103], and the
Research on the use of digital twins in product design is growing; application of data mining techniques for predicting the design change
however, it is still in its early stages. propagation [104].

10
D. Vlah et al. Advanced Engineering Informatics 54 (2022) 101774

Table 4 Table 4 (continued )

The twenty most cited articles for data-driven engineering design. Author Year Source Cit. Title
Author Year Source Cit. Title
Attractiveness of a
1 Tao F; Cheng J; Qi 2018 Int. J. Adv. 443 Digital Twin-Driven Passenger Car Form
Q; Zhang M; Zhang Manuf. Product Design, Aimed at Young
H; Sui F [80] Technol. Manufacturing and Consumers
Service with Big Data 18 Aguwa C; 2012 Expert 43 Voice Of the
2 Kusiak A [99] 2018 Int. J. 316 Smart Manufacturing Monplaisir L; Turgut Syst. Appl. Customer: Customer
Prod. Res. O [111] Satisfaction Ratio
3 Wang X; Feng C 2002 Int. J. Adv. 248 Development Of Based Analysis
[100] Manuf. Empirical Models for 19 Bae J; Kim J [112] 2011 Expert 40 Product Development
Technol. Surface Roughness Syst. Appl. with Data Mining
Prediction in Finish Techniques: A Case on
Turning Design of Digital
4 Qi Q; Tao F [101] 2018 IEEE 230 Digital Twin and Big Camera
Access Data Towards Smart 20 Lee C; Song B; Park 2012 Expert 37 Design Of Convergent
Manufacturing and Y [113] Syst. Appl. Product Concepts
Industry 40 360 Based on
Degree Comparison Functionality an
5 Agard B; Kusiak A 2004 Int. J. 136 Data Mining Based Association Rule
[53] Prod. Res. Methodology for the Mining and Decision
Design of Product Tree Approach
Families
6 Zhan J; Loh H; Liu Y 2009 Expert 134 Gather Customer
[102] Syst. Appl. Concerns from Online In addition, we also considered the association between the number
Product Reviews A of citations that the publication received and the impact factor of the
Text Summarization
journal in which the publication was published. We computed Pearson’s
Approach
7 Tao F; Sui F; Liu A; 2019 Int. J. 103 Digital Twin Driven correlation coefficient between the column CIT from Table 4 and the
Qi Q; Zhang M; Song Prod. Res. Product Design columns of corresponding impact factors (not reported in the paper),
B; Guo Z; Lu S; Nee Framework which is equal to r = 0.29. Association is, therefore, positive, although
A [77] not significant.
8 Ahmed S; Kim S; 2007 J. Comput. 92 A Methodology for
Wallace K [103] Inf. Sci. Creating Ontologies
Eng. for Engineering 3.4. RQ4: BDA applications in product development
Design
9 Pasqual M; De W O 2012 Res. Eng. 80 Multilayer Network
[104] Des. Model for Analysis This section focuses on how BDA applications enter specific product
and Management of development activities. The number of articles related to DDED extrac­
Change Propagation ted for different product development activities is shown in Fig. 8. A
10 Wu D; Thames J; 2013 J. Comput. 74 Enhancing The total of 366 articles were screened, and the articles were categorised
Rosen D; Schaefer D Inf. Sci. Product Realization
[105] Eng. Process with Cloud
based on keywords from the article titles using text filtering features.
Based Design and Consequently, a total of 249 articles were assigned into categories;
Manufacturing however, 51 of them were suitable for the two or more categories
Systems considering the keywords. In Fig. 8, these articles were included in all
11 Wang K [106] 2011 Expert 58 A Hybrid Kansei
categories, where they were automatically sorted. Owing to the lack of
Syst. Appl. Engineering Design
Expert System Based distinct borderlines between the product development phases and ac­
on Grey System tivities, which also occasionally overlap, such articles were assigned to
Theory and Support more categories. Thus, in total, 117 out of 366 relevant articles obtained
Vector Regression from the database were left uncategorized by product development ac­
12 Kusiak A [107] 2007 Comput.- 55 Innovation The Living
Aided Des. Laboratory
tivities. These include broader review articles, studies focused on edu­
Appl. Perspective cation, extraction of overall design rationale and heuristics, design for
13 Darses F; Wolff M 2006 Appl. 54 How Do Designers sustainability, design collaboration studies, process mining, and other
[108] Ergon. Represent to topics, which do not answer our research question.
Themselves the Users
The articles that were categorised into different product develop­
Needs
14 Jin J; Liu Y; Ji P; Liu 2016 Int. J. 53 Understanding Big ment activities were further sorted according to the BDA techniques
H [76] Prod. Res. Consumer Opinion used. The automatic search of keywords related to BDA techniques was
Data for Market performed on the article titles and abstracts. Fig. 9 presents the mapping
Driven Product of BDA techniques to product development activities, for which appli­
Design
15 Ravikumar S; 2011 Expert 51 Machine Learning
cation examples were found in the articles. As all the articles did not
Ramachandran K; Syst. Appl. Approach for employ BDA techniques stated in the title or abstract the mapping
Sugumaran V [109] Automated Visual contained fewer articles than the categorisation by product development
Inspection of Machine activities. A brief description of the main applications of specific BDA
Components
techniques in different activities is provided below.
16 Tuarob S; Tucker C 2015 J. Comput. 49 Quantifying Product
[95] Inf. Sci. Favourability and
Eng. Extracting Notable 3.4.1. Market research
Product Features At the beginning of the product development process, market op­
Using Large Scale portunities and market segments must be defined, along with new
Social Media Data
17 Chang H; Lai H; 2007 Int. J. Ind. 44 A Measurement Scale
technologies to be considered to define the product requirements and
Chang Y [110] Ergon. for Evaluating the positioning among the potential users, to clarify the design task [60].
Regression techniques can be used to predict the price or future product
demand based on users’ preferences toward design features in existing

11
D. Vlah et al. Advanced Engineering Informatics 54 (2022) 101774

Fig. 8. Distribution of articles related to BDA applications in specific product development activities.

Fig. 9. BDA techniques applied in different product development activities. The size of the bubble on each intersection shows the number of identified articles.

product designs [115,116]. Christian et al. [117] applied text mining to 3.4.2. User research
evaluate the market favourability of new, creative product ideas. In Identification of users, their needs, and preferences can be enhanced
addition, BDA technologies can also be utilised to assist designers with by applying data-driven approaches that aid in overcoming certain is­
problem exploration [118] and with the discovery of new technologies, sues of conventional product design, which rely on surveys and user
which could improve designs [84,119]. experiments [120]. Mining user and product data from web sources such
as e-commerce online reviews and social media has garnered research
attention because of the valuable information contained regarding

12
D. Vlah et al. Advanced Engineering Informatics 54 (2022) 101774

sentiments and expectations that users express about products without similarity between concepts and clustering. In addition, data mining can
restrictions [73,121]. However, the use of online data sources results in also be used to aid concept selection and evaluation [147,148]. The
new challenges, primarily identifying and extracting semantic infor­ proposed approaches accounted for different evaluation aspects such as
mation from unstructured text, but also addressing the existence of novelty [149], level of detail, and feasibility [148].
implicit data such as sarcasm, which can be misleading for creating
predictive models [24]. Common applications of text mining and natural 3.4.6. Embodiment design
language processing are opinion mining [22,122,123] and affective During embodiment design designers must determine the overall
design [21,75,124], which are often approached using the Kansei en­ layout design (general arrangement and spatial compatibility) and
gineering method [125,126] and sentiment analysis [127–129]. Certain preliminary form designs (component shapes and materials) adhering to
examples include discovering relations between product features and the technological and economic criteria [5]. Examples of BDA applica­
opinions [130], attribute ranking [131], customer needs and prefer­ tions include automatic shape design space exploration, as well as the
ences elicitation [132,133], extracting customer affordance levels and integration of BDA into topology optimisation [150] and generative
defining usage conditions [134]. Chen et al. [135] explored different design systems [151]. Classification algorithms are often used to retrieve
techniques, such as regression, ANNs, and random forests, to achieve potential design alternatives from a set of designs. Graening et al. [152]
user preference extraction. In [89], SVM and deep learning were applied presented a holistic approach using DTs to realize shape mining, which
to realize affective attribute classification. Further, other studies pro­ combines data from different design representations, tools, and methods
posed the use of association rules and decision trees to define user for the evaluation of design performance. Zhang et al. [153] applied the
preferences on product features [73,112,121], whereas a study [136] modified SVM technique to historical product development data to
presented a method for customer segmentation using clustering. In predict the performance values for design alternatives, whereas Dasari
addition, BDA-based models are often used to build user-centred design et al. [154] used random forest surrogate models for prediction. Further,
recommendation systems. Angrish et al. [155] proposed the use of 3D CAD data and applied CNNs
for shape classification and retrieval.
3.4.3. Definition of requirements
Customer needs are typically expressed in a form of the ‘language of 3.4.7. CAD and CAE
the customer’ and must be translated into design requirements (speci­ CAD and CAE are essential tools employed for the representation and
fications), which define precisely, in technical terms, what the product validation of designs. CAD refers to the tools that enable the creation of
must do [60]. Two of the identified studies used ontologies to enable 3D product models, whereas CAE encompasses a group of tools used for
semantic processing to support automated requirements elicitation conducting engineering simulations on virtual product prototypes to
[82,137]. Moreover, text mining of social media [24] and online reviews obtain a better understanding of design performance. However, CAD
[131] can be used to extract and prioritise the requirements of product and CAE encounter a significant challenge of extraction and conversion
features that are crucial for design. Jiao et al. [138] tested different of relevant data into usable formats [156]. The potential use of neural
classification techniques for mapping customer requirements to product networks for CAD automation has been described by Kries et al. [157].
configurations using data obtained from e-commerce platforms and The data extracted from the CAD models can be reused for future de­
customer interviews. Bigorra et al. [139] enhanced the house of quality signs. Bickel et al. [158] applied classification and clustering techniques
(HoQ) with BDA to link design requirements with customer needs and to assess the geometric similarity of the models, whereas Vasantha [159]
product use, whereas Li and Wang [140] used data mining to auto­ used association rules to identify potentially substitutable geometrical
matically identify the relationship between the needs and design spec­ features. Furthermore, CAD event data can aid in identifying meaningful
ifications in HoQ. design operations [160]. The existing simulation data can be utilised to
support the new simulation case setup [90] and to examine the plausi­
3.4.4. Function analysis bility of the simulation results [161]. In addition, such data can be used
In the design process, downstream information regarding product to guide design optimisation and improve the optimum search efficiency
and component functions is commonly used to facilitate ideation and [162]. Furthermore, for large datasets of simulation results, dimen­
concept generation of analogous products [141]. Function analysis re­ sionality reduction methods can be used to extract the influential
fers to the process of defining and exploring product functions. In several simulation parameters [163].
studies, association rule mining was applied to identify the functions and
critical product features [113] and to consequently map the specific 3.4.8. Product platform assessment
functions with adequate technologies [142] or product components Product family design and product configuration based on BDA
[143]. In this activity, the decision trees can be used to extract the re­ technology have been identified as intelligent and automated means to
lations between the components and functions [141], while also sup­ improve product development efficiency [164]. Association rule mining
porting the recommendation of specifications with respect to target is a commonly used technique for linking customer needs and associated
functions [113]. To define the function-component relations, the au­ product modules and creating classifiers for the recommendation of the
thors proposed using existing companies’ technical documents and best configurations [164,165]. In addition, clustering can be used to
design repositories. Kang and Tucker [144] applied text mining to identify customer segments [165], product family modules [166], define
extensive product specification data to automate the quantification of component size ranges according to customer requests [167], and group
functional interactions between modules. Zhang et al. [145] developed a product variants according to production requirements [168]. Further,
recommender system for new product functions by mining e-commerce dimensionality reduction techniques can help reduce the number of
data. customer needs or product attributes [164]. Tucker and Kim [96] used
DTs to create optimal portfolios according to users’ price and perfor­
3.4.5. Principal concept development mance expectations, and profit. In most studies, historical customer
Product concept development considers the search for solutions and request and survey data were used.
working principles for the defined functions and their evaluation. The
idea generation process can be aided by techniques such as text mining 3.4.9. Detailing
and natural language processing (NLP)-based networks capable of Detailed design includes the creation of products’ technical docu­
extracting relevant concepts from large databases such as patents [87], mentation considering manufacturing and assembly (assigning toler­
web documents [83], and social networks [146], or data collected ances, creating a bill of materials (BOM), etc.). In this stage, many
through crowdsourcing. The common approaches are measuring the processes can be automatised using data from previous designs.

13
D. Vlah et al. Advanced Engineering Informatics 54 (2022) 101774

Designers can utilise text mining and clustering techniques to identify main research questions were posed to explore various aspects of state-
similarities in historical BOM structures to address the variability in of-the-art research in DDED.
product platform design [169,170]. Zhao et al. [171] applied SVR to First, the yearly dynamics of the articles were explored, and a trend
optimise the product tolerancing system, whereas Yu et al. [172] sug­ of rapid growth was recognised. In the past two decades, more than 60 %
gested an association rule-based model for the recommendation of en­ of all articles were published only in the last five years, indicating that
gineering note items. Moreover, through the utilisation of BDA tools, the DDED research will continue to develop and expand. Owing to the
planning and design of assembly processes can be improved, thereby recent advancements in data collection and processing technologies,
resulting in time and cost reductions [173,174]. along with the democratisation of tools and methods that enable BDA,
such results were expected.
3.4.10. Testing and refinement Thereafter, we investigated the primary topics of interest in DDED
BDA can be implemented in testing activities to enhance the product through the analysis of frequent keywords in published papers and a
performance analysis and user behaviour assessment. Several authors keyword co-occurrence network. A range of topics were identified with
have investigated the use of product prototypes for data collection and customer opinion mining, optimisation, creativity, and design knowl­
proposed big-data-based tools for usability testing [175,176]. Based on edge management identified as the most common. Additional topics
the experimental data acquired during testing, BDA models can be built frequently occurring in the articles were related to the growing number
to predict and optimise product performance. However, when imple­ of connected products, which included topics such as product-service
menting refinements, the influence of engineering changes must be systems, human engineering, and smart manufacturing.
estimated. NLP on company documents was suggested by Arnarsson Furthermore, the most prominent authors and highly cited papers
et al. [177] to predict the change propagation, whereas Zhang et al. were explored. It was found that research groups in DDED were quite
[178] proposed a new automated approach for change recommendation dispersed and disconnected. This emphasises the need to improve
for connected products. collaboration between groups to promote the development and growth
of the research field. The leading research groups with the most publi­
3.4.11. Production cations and connections include academics from China, Singapore, the
According to Kusiak [99], smart manufacturing is a new form of United States of America, and Germany. Their researched topics cover
production that integrates current manufacturing assets with sensors, mostly customer data mining for affective design, data-supported con­
computing platforms, communication technology, control, simulation, ceptual design, product-service systems design, digital twins, and design
data-intensive modelling, and predictive engineering. Major concepts in knowledge extraction (early product development phases and products
this area include cyber-physical systems and digital twins, cloud-based in use). The top-cited publications provided comparable results, high­
design, manufacturing, servitisation in product-service systems, and lighting the need for exploration and integration of design data collected
human-robot collaboration [179]. BDA plays a significant role in the through diverse sources, such as online customer reviews, patents,
realisation of smart manufacturing systems, also affecting product previous design projects, historical data, and sensors.
design. Exemplary applications include production cost assessment Advantages of using various data sources are many: combining in­
[180,181], visual inspection and quality control [109,182,183], product sights from different data can help to understand the users and their
manufacturability assessment [184], fault prediction [185], and the use behaviours, quantitatively describe their preferences, help defining re­
of deep learning for computer vision [186]. Furthermore, the articles quirements, support ideation and creation of various design embodi­
highlight various approaches to realise machining parameter prediction, ments with respect to feasibility and performance, search for
for example, classification and regression techniques used for surface multicriteria optimal solutions and adjust designs for production.
roughness prediction in turning [100], applications in welding [187], Analysis and utilization of the product data provides a clearer,
sheet metal forming [96], and additive manufacturing [188]. Romeo comprehensive characterization of the product, user behaviour and their
et al. [189] proposed a method using decision trees and KNNs to predict interaction.
the machine design specification data. By analysing the published works, several future perspectives were
extracted:
3.4.12. Product use
Advances in information and communication technologies have • Research findings indicate that the BDA would become useful in
boosted the development of connected products, enabling more accurate conceptual design phase to enhance creativity and support ideation
recording of products during use and their timely responsiveness to and concept generation by recommending the potential functions
users’ actions. Consequently, smart product-service systems have and concept solutions to designers.
emerged, which enable designers to gain better insight into how users • To support the further steps of the product development process, we
interact with the product, and the users are provided with improved expect the development of BDA-supported product design, simula­
services in return. Certain studies proposed methods that utilise the data tion, and management tools, with advanced recommendation and
collected from sensors [190,191], or combining both online and offline design assistance capabilities. Embodiment design and succeeding
user-generated and product-sensed data [83], to optimise the design and product development activities require more specific, numerical, and
provide new services to users. Song et al. employed text mining on categorical engineering data, which relate to techniques such as
historical service documents to develop a service evolution map regression and classification. These trends can already be noticed in
considering product elements and service concepts. In contrast to approaches such as generative design or in the use of machine
retrospective product use analysis, [192] proposed a method for assist­ learning for multi-parameter optimisation, but new development in
ing the anticipation of future product use. Furthermore, applying that field can also be expected.
customer clustering and analysing similar usage scenarios, He et al. • The development of new data acquisition and management tools and
[193] proposed a framework for developing user choice prediction algorithms will open the possibility to exploit new data sources. They
models, which can help designers with strategic decision-making for could incorporate different means of artificial intelligence, which
new product design. would give sense to the collected data and thus help decision-making
in different stages of development processes. In addition to the rapid
4. Conclusion analysis and interpretation capabilities, such intelligent systems
would assist designers by detecting human errors and providing
In this paper, a scientometric approach to the systematic review of appropriate proposals, as well as discovering new and unexplored
data-driven engineering design (DDED) research was presented. Four connections between the design items.

14
D. Vlah et al. Advanced Engineering Informatics 54 (2022) 101774

• Greater availability and low price of sensors have facilitated the rise [4] S. Ren, Y. Zhang, Y. Liu, T. Sakao, D. Huisingh, C.M.V.B. Almeida,
A comprehensive review of big data analytics throughout product lifecycle to
of connected products, which has directed research in exploration of
support sustainable smart manufacturing: A framework, challenges and future
data collected during product use. Most present works in this area research directions, J. Clean. Prod. 210 (2019) 1343–1365, https://doi.org/
introduce methods for identifying useful data and design potential. 10.1016/J.JCLEPRO.2018.11.025.
However, implementation strategies and exemplary case studies [5] G. Pahl, W. Beitz, J. Feldhusen, K.-H. Grote, Engineering Design: A systematic
approach, 3rd ed., Springer-Verlag, London, 2007. https://doi.org/10.1007/978-
remain scarce. Further implementation of small-scale sensors on full 1-84628-319-2.
range of different products will provide an opportunity for the [6] T. Sakao, What is PSS design? – Explained with two industrial cases, Procedia -
collection of new data that will monitor and predict user engagement Soc, Behav. Sci. 25 (2011) 403–407, https://doi.org/10.1016/J.
SBSPRO.2012.02.051.
with the products. These data could be used for improving the next [7] S.A. Chowdhery, M. Bertoni, J. Wall, T. Larsson, A data-driven design framework
generation of products and new designs. for early stage PSS design exploration, Des, Sci, 2020.
• Underexplored opportunities are also found in crowdsourcing as a [8] M.K.D. Haffey, A.H.B. Duffy, Knowledge Discovery and Data Mining within a
Design Environment, in: From Knowl. Intensive CAD to Knowl. Intensive Eng.,
method for data collection, as a part of collaborative and integrated Springer US, 2002: pp. 59–74. https://doi.org/10.1007/978-0-387-35494-1_5.
product development. This will provide new insights about products, [9] D. Sriram, G. Stephanopoulos, R. Logcher, D. Gossard, N. Groleau, D. Serrano,
such as information extracted from previous designs and crowd­ D. Navinchandra, Knowledge-Based System Applications in Engineering Design:
Research at MIT, AI Mag. 10 (1989) 79, https://doi.org/10.1609/AIMAG.
sourced design ideas. V10I3.758.
• All above mentioned perspectives would eventually lead to new [10] S. Kleiner, R. Anderl, R. Gräb, A collaborative design system for product data
business and research opportunities, like creation of new databases integration, J. Eng. Des. 14 (2003) 421–428, https://doi.org/10.1080/
09544820310001606902.
(e.g., from web scraping, from company internal data, patents etc.)
[11] E. Begoli, J. Horey, Design principles for effective knowledge discovery from big
and data-based tools supporting different design tasks. data, Proc. 2012 Jt. Work. Conf. Softw. Archit. 6th Eur. Conf. Softw. Archit.
WICSA/ECSA 2012. (2012) 215–218. https://doi.org/10.1109/WICSA-
Of course, the path to these goals is not straightforward and several ECSA.212.32.
[12] M. Rogalewicz, R. Sika, Methodologies of knowledge discovery from data and
challenges need to be addressed. Before the implementation of DDED data mining methods in mechanical engineering, Manag. Prod. Eng. Rev. 7
can take broader place, significant efforts need to be invested into the (2016) 97–108, https://doi.org/10.1515/mper-2016-0040.
next three R&D domains: (i) into the development of tools for easy [13] J.H. Gennari, M.A. Musen, R.W. Fergerson, W.E. Grosso, M. Crubezy, H. Eriksson,
N.F. Noy, S.W. Tu, The evolution of Protégé: an environment for knowledge-
collecting and interpretation of relevant data and (ii) into the systematic based systems development, Int. J. Hum. Comput. Stud. 58 (2003) 89–123,
and methodological integration of data collection and their analyses into https://doi.org/10.1016/S1071-5819(02)00127-1.
the design process. Based on that, (iii) a formulation of new data- [14] T.H. Nguyen, B.Q. Vo, M. Lumpe, J. Grundy, KBRE: a framework for knowledge-
based requirements engineering, Softw. Qual. J. 2013 221. 22 (2013) 87–119.
supported design processes will be needed as well as the development https://doi.org/10.1007/S11219-013-9202-6.
of cocreation and design support tools that efficiently utilize the solu­ [15] G. La Rocca, Knowledge based engineering: Between AI and CAD. Review of a
tions to beforementioned challenges. language based technology to support engineering design, Adv. Eng. Informatics.
26 (2012) 159–179, https://doi.org/10.1016/J.AEI.2012.02.002.
Across all these domains of challenges there remains an open ques­ [16] S. Preston, C. Chapman, M. Pinfold, G. Smith, Knowledge acquisition for
tion, how to acquire and utilize data from diverse sources e.g., user re­ knowledge-based engineering systems, Int. J. Inf. Technol. Manag. 4 (2005) 1–11,
sponses, computational studies, models, testing, which are usually https://doi.org/10.1504/IJITM.2005.006401.
[17] D. Baxter, J. Gao, Development of a Process Based Data Driven Engineering
delivered in different forms, such as textual, numerical or graphical.
Design Knowledge Reuse System, Comput. Aided. Des. Appl. 3 (2006) 109–117,
Important data-science question is, how to extract high-quality and https://doi.org/10.1080/16864360.2006.10738447.
credible data from the mass of data and how to exclude inadequate in­ [18] Q. Chen, G. Tan, J. Han, W. He, Design and implementation of an intelligent
formation and thereby derive correct conclusions, results and design knowledge discovery system, Proc. 11th Jt. Int. Comput. Conf. JICC (2005.
(2005)) 482–487, https://doi.org/10.1142/9789812701534_0109.
decisions. We believe that with the ever-increasing amount of available [19] C. Liu, X. Chen, Data-driven design paradigm in engineering problems:, in: Proc.
data from various sources this will become even more increasing chal­ Inst. Mech. Eng., SAGE Publications, London, 2016: pp. 1522–1534. https://doi.
lenge in the future. org/10.1177/0954410016653502.
[20] Y. Feng, Y. Zhao, H. Zheng, Z. Li, J. Tan, Data-driven product design toward
intelligent manufacturing: A review, 172988142091125, Int. J. Adv. Robot. Syst.
17 (2020), https://doi.org/10.1177/1729881420911257.
Declaration of Competing Interest [21] Z. Li, Z.G. Tian, J.W. Wang, W.M. Wang, Extraction of affective responses from
customer reviews: an opinion mining and machine learning approach, Int. J.
Comput. Integr. Manuf. 33 (2019) 670–685, https://doi.org/10.1080/
The authors declare that they have no known competing financial
0951192X.2019.1571240.
interests or personal relationships that could have appeared to influence [22] J. Jin, Y. Liu, P. Ji, C.K. Kwong, Review on recent advances in information mining
the work reported in this paper. from big consumer opinion data for product design, J. Comput. Inf. Sci. Eng. 19
(2019), https://doi.org/10.1115/1.4041087.
[23] W.M. Wang, Z. Li, L. Liu, Z.G. Tian, E. Tsui, Mining of affective responses and
Data availability affective intentions of products from unstructured text, J. Eng. Des. 29 (2018)
404–429, https://doi.org/10.1080/09544828.2018.1448054.
Data will be made available on request. [24] S. Tuarob, S. Lim, C.S. Tucker, Automated Discovery of Product Feature
Inferences Within Large-Scale Implicit Social Media Data, J. Comput. Inf. Sci.
Eng. 18 (2018), https://doi.org/10.1115/1.4039432.
Acknowledgements [25] M. Preidel, W.M. Wang, K. Exner, R. Stark, Knowledge in engineering design: A
systematic literature review on artifacts and it systems, Proc. Int. Des. Conf. Des. 1
(2018) 881–892. https://doi.org/10.21278/IDC.2018.0220.
We thank the lectors from the Elsevier Language Editing Service for [26] S. Bickel, C. Sauer, B. Schleich, S. Wartzack, Comparing CAD part models for
the proof-reading and useful suggestions. This research was co-funded geometrical similarity: A concept using machine learning algorithms, Procedia
by the Slovenian Research Agency (Grant Nos. P2-0425 and J5-2552). CIRP. 96 (2021) 133–138, https://doi.org/10.1016/J.PROCIR.2021.01.065.
[27] T.C. Spruegel, S. Bickel, B. Schleich, S. Wartzack, Approach and application to
transfer heterogeneous simulation data from finite element analysis to neural
References networks, J. Comput. Des. Eng. 8 (2021) 298–315, https://doi.org/10.1093/jcde/
qwaa079.
[28] T. Stürmlinger, B. Gladysz, M. Strauch, A. Albers, Design with Industry 4.0 –
[1] Z. Sun, L. Sun, K. Strang, Big Data Analytics Services for Enhancing Business
Priorization of Sensor data for a smart data driven Product Development Process,
Intelligence, J. Comput. Inf. Syst. 58 (2018) 162–169, https://doi.org/10.1080/
in, Proc. TMCE, Las Palmas de Gran Canaria (2018).
08874417.2016.1220239.
[29] J. Wilberg, F. Schafer, P. Kandlbinder, C. Hollauer, M. Omer, U. Lindemann, Data
[2] Z. Sun, Y. Huo, The Spectrum of Big Data Analytics, J. Comput. Inf. Syst. 61
analytics in product development: Implications from expert interviews, IEEE Int.
(2021) 154–162, https://doi.org/10.1080/08874417.2019.1571456.
Conf. Ind. Eng. Eng. Manag. (2017-December (2018)) 818–822, https://doi.org/
[3] M. Cantamessa, F. Montagna, S. Altavilla, A. Casagrande-Seretti, Data-driven
10.1109/IEEM.2017.8290005.
design: The new challenges of digitalization on product design and development,
Des. Sci. 6 (2020), https://doi.org/10.1017/dsj.2020.25.

15
D. Vlah et al. Advanced Engineering Informatics 54 (2022) 101774

[30] Í.Ö. Arnarsson, E. Gustavsson, J. Malmqvist, M. Jirstrand, Design analytics is the M. Grimshaw, A. Hróbjartsson, M.M. Lalu, T. Li, E.W. Loder, E. Mayo-Wilson,
answer, but what questions would product developers like to have answered?, in, S. McDonald, L.A. McGuinness, L.A. Stewart, J. Thomas, A.C. Tricco, V.A. Welch,
Proc. 21st Int. Conf. Eng. Des. ICED 17 (2017) Vancouver. P. Whiting, D. Moher, The PRISMA 2020 statement: an updated guideline for
[31] A. Henry, L. Stieglitz, An examination of systematic reviews in the engineering reporting systematic reviews, BMJ 372 (2021), https://doi.org/10.1136/BMJ.
literature, ASEE Annu, Conf. Expo. Conf. Proc. (2020-June (2020).). https://doi. N71.
org/10.18260/1-2–34121. [58] A. Albers, N. Bursac, V. Butenko, F. Marthaler, Q. Zhang, The world is complex
[32] J. Trauer, S. Schweigert-Recksiek, L.O. Okamoto, K. Spreitzer, M. Mörtl, M. therefore our research is needed - Using data mining for literature reviews, Proc.
Zimmermann, Data-driven engineering definitions and insights from an industrial Int. Des. Conf. Des. 1 (2018) 1–12. https://doi.org/10.21278/IDC.2018.0470.
case study for a new approach in technical product development, Proc. Nord. [59] Y. Chen, W. Wang, Z. Liu, Keyword-based search and exploration on databases,
2020 Conf. Nord. 2020. (2020). https://doi.org/10.35199/ Proc. - Int. Conf. Data Eng. (2011) 1380–1383, https://doi.org/10.1109/
NORDDESIGN2020.46. ICDE.2011.5767958.
[33] D. Van Horn, A. Olewnik, K. Lewis, Design analytics: Capturing, understanding, [60] K.T. Ulrich, S.D. Eppinger, M.C. Yang, Product Design and Development, 7th ed.,
and meeting customer needs using big data, Proc. ASME Des. Eng. Tech. Conf. 7 McGraw-Hill, New York City, 2019.
(2012) 863–875, https://doi.org/10.1115/DETC2012-71038. [61] R.G. Cooper, The Stage-Gate Idea to Launch System, Wiley Int. Encycl. Mark.
[34] F. Chiarello, P. Belingheri, G. Fantoni, Data science for engineering design: State (2010), https://doi.org/10.1002/9781444316568.WIEM05014.
of the art and future directions, Comput. Ind. 129 (2021), 103447, https://doi. [62] Verein Deutsche Ingenieure, VDI 2221 - Entwicklung technischer Produkte und
org/10.1016/J.COMPIND.2021.103447. Systeme, Beuth-Verlag, 2019. https://www.beuth.de/en/technical-rule/vdi-
[35] A. Bertoni, Data-driven design in concept development: systematic review and 2221-blatt-1/311603768 (accessed October 14, 2021).
missed opportunities, in: Proc. Des. Soc. Des. Conf., Cambridge University Press [63] D. Dhall, R. Kaur, M. Juneja, Machine Learning: A Review of the Algorithms and
(CUP), 2020: pp. 101–110. https://doi.org/10.1017/dsd.2020.4. Its Applications, Lect. Notes Electr. Eng. 597 (2020) 47–63, https://doi.org/
[36] S.S. Shabestari, M. Herzog, B. Bender, A Survey on the Applications of Machine 10.1007/978-3-030-29407-6_5.
Learning in the Early Phases of Product Development, Proc. Des. Soc. Int. Conf. [64] EMC Education Series, Data science and big data analytics : discovering,
Eng. Des. 1 (2019) 2437–2446, https://doi.org/10.1017/DSI.2019.250. analyzing, visualizing and presenting data, 1st ed., John Wiley & Sons,
[37] F. Montagna, M. Cantamessa, Unpacking the innovation toolbox for design Indianapolis, IN, USA, 2015.
research and practice, Des. Sci. 5 (2019), https://doi.org/10.1017/DSJ.2019.3. [65] J. Han, M. Kamber, J. Pei, Data mining : Concepts and Techniques, 3rd ed.,
[38] M. Meyer, I. Wiederkehr, C. Koldewey, R. Dumitrescu, Understanding Usage Morgan Kaufmann, Waltham, MA, USA, 2012.
Data-driven Product Planning: a Systematic Literature Review, in: Proc. Int. Conf. [66] M.M. Hasan, J. Popp, J. Oláh, Current landscape and influence of big data on
Eng. Des., Cambridge University Press (CUP), Gothenburg, Sweden, 2021: pp. finance, J. Big Data 2020 71. 7 (2020) 1–17. https://doi.org/10.1186/S40537-
3289–3298. https://doi.org/10.1017/PDS.2021.590. 020-00291-Z.
[39] K. Gorkovenko, D.J. Burnett, J.K. Thorp, D. Richards, D. Murray-Rust, Exploring [67] R.A. Jafar, A. Sajid, R. Talib, M. Niaz, A Review on the Role of Big Data in
the Future of Data-Driven Product Design, in: Conf. Hum. Factors Comput. Syst. - Business, Int. J. Comput. Sci. Mob. Comput. 3 (2014) 446–453.
Proc., Association for Computing Machinery, 2020. https://doi.org/10.1145/ [68] M. Blackburn, J. Alexander, J. Legan, D. Klabjan, Big data and the future of R&D
3313831.3376560. management: the rise of big data and big data analytics will have significant
[40] H. Zheng, Y. Feng, Y. Gao, J. Tan, A Robust Predicted Performance Analysis implications for R&D and innovation management in the next decade, Res
Approach for Data-Driven Product Development in the Industrial Internet of Technol Manag. 60 (2017) 43–51, https://doi.org/10.1080/
Things, Sensors 2018, Vol. 18, Page 2871. 18 (2018) 2871. https://doi.org/ 08956308.2017.1348135.
10.3390/S18092871. [69] K.Z.A.C.S.Y.S.E. fezazi A Belhadi, Understanding big data analytics for
[41] L. Wang, Z. Liu, A. Liu, F. Tao, Artificial intelligence in product lifecycle manufacturing processes: insights from literature review and multiple case
management, Int. J. Adv. Manuf. Technol. 1143 (114) (2021) 771–796, https:// studies, Comput Ind Eng. 137 (2019) 106099. https://doi.org/10.1016/j.
doi.org/10.1007/S00170-021-06882-1. cie.2019.106099.
[42] Y. Li, U. Roy, J.S. Saltz, Towards an integrated process model for new product [70] A. Amado, P. Cortez, P. Rita, S. Moro, Research trends on Big Data in Marketing:
development with data-driven features (NPD3), Res. Eng. Des. 302 (30) (2019) A text mining and topic modeling based literature analysis, Eur. Res. Manag. Bus.
271–289, https://doi.org/10.1007/S00163-019-00308-6. Econ. 24 (2018) 1–7, https://doi.org/10.1016/J.IEDEEN.2017.06.002.
[43] D. Ullman, The Mechanical Design Process, 4th ed., McGraw-Hill Education, New [71] Q. Yang, Machine Learning as a UX Design Material: How Can We Imagine
York, NY, USA, 2010. Beyond Automation, Recommenders, and Reminders?, in: AAAI Spring Symp.
[44] A. Bernard, E. Coatanea, F. Christophe, F. Laroche, Design: A Key Stage of Product Ser. User Exp. Artif. Intell., Palo Alto, CA, 2018.
Lifecycle, Procedia CIRP. 21 (2014) 3–9, https://doi.org/10.1016/J. [72] S.S. Iyer, K.I. Lakhtaria, Practical Evaluation and Comparative Study of Big Data
PROCIR.2014.06.146. Analytical Tools, in: Int. Conf. ’Recent Trends IT Innov., Bangalore, 2017.
[45] R. van Dinter, B. Tekinerdogan, C. Catal, Automation of systematic literature [73] H. Jiang, C.K. Kwong, W.Y. Park, K.M. Yu, A multi-objective PSO approach of
reviews: A systematic literature review, Inf. Softw. Technol. 136 (2021), 106589, mining association rules for affective design based on online customer reviews,
https://doi.org/10.1016/j.infsof.2021.106589. J. Eng. Des. 29 (2018) 381–403, https://doi.org/10.1080/
[46] K.R. Felizardo, J.C. Carver, Automating Systematic Literature Review, Contemp. 09544828.2018.1475629.
Empir. Methods Softw. Eng. (2020) 327–355, https://doi.org/10.1007/978-3- [74] H. Jiang, C.K. Kwong, G.E. Okudan Kremer, W.Y. Park, Dynamic modelling of
030-32489-6_12. customer preferences for product design using DENFIS and opinion mining, Adv.
[47] G. Lame, Systematic Literature Reviews: An Introduction, Proc. Des. Soc. Int. Eng. Informatics. 42 (2019), 100969, https://doi.org/10.1016/J.
Conf. Eng. Des. 1 (2019) 1633–1642, https://doi.org/10.1017/DSI.2019.169. AEI.2019.100969.
[48] P.V. Torres-Carrion, C.S. Gonzalez-Gonzalez, S. Aciar, G. Rodriguez-Morales, [75] K.Y. Chan, C.K. Kwong, P. Wongthongtham, H. Jiang, C.K.Y. Fung, B. Abu-Salih,
Methodology for systematic literature review applied to engineering and Z. Liu, T.C. Wong, P. Jain, Affective design using machine learning: a survey and
education, IEEE Glob. Eng. Educ. Conf. EDUCON. 2018-April (2018) 1364–1373. its prospect of conjoining big data, Int. J. Comput. Integr. Manuf. 33 (2020)
https://doi.org/10.1109/EDUCON.2018.8363388. 645–669, https://doi.org/10.1080/0951192X.2018.1526412.
[49] L. Rabhi, N. Falih, A. Afraites, B. Bouikhalene, Big Data Approach and its [76] J. Jin, Y. Liu, P. Ji, H. Liu, Understanding big consumer opinion data for market-
applications in Various Fields: Review, in, Procedia Comput. Sci., Elsevier B.V. driven product design, Int. J. Prod. Res. 54 (2016) 3019–3041, https://doi.org/
(2019) 599–605, https://doi.org/10.1016/j.procs.2019.08.084. 10.1080/00207543.2016.1154208.
[50] M.S. Ali, M. Ali Babar, L. Chen, K.J. Stol, A systematic review of comparative [77] F. Tao, F. Sui, A. Liu, Q. Qi, M. Zhang, B. Song, Z. Guo, S.C.Y. Lu, A.Y.C. Nee,
evidence of aspect-oriented programming, Inf. Softw. Technol. 52 (2010) Digital twin-driven product design framework, Int. J. Prod. Res. 57 (2019)
871–887, https://doi.org/10.1016/J.INFSOF.2010.05.003. 3935–3953, https://doi.org/10.1080/00207543.2018.1443229.
[51] J. Mingers, L. Leydesdorff, A review of theory and practice in scientometrics, Eur. [78] Y. Liu, A. Soroka, L. Han, J. Jian, M. Tang, Cloud-based big data analytics for
J. Oper. Res. 246 (2015) 1–19, https://doi.org/10.1016/J.EJOR.2015.04.002. customer insight-driven design innovation in SMEs, Int. J. Inf. Manage. 51
[52] M.K.D. Haffey, A.H.B. Duffy, Knowledge Discovery and Data Mining within a (2020), https://doi.org/10.1016/J.IJINFOMGT.2019.11.002.
Design Environment, From Knowl. Intensive CAD to Knowl, Intensive Eng. (2000) [79] Q. Liu, K. Wang, Y. Li, Y. Liu, Data-Driven Concept Network for Inspiring
59–74, https://doi.org/10.1007/978-0-387-35494-1_5. Designers’ Idea Generation, J. Comput. Inf. Sci. Eng. 20 (2020), https://doi.org/
[53] B. Agard, A. Kusiak, Data-mining-based methodology for the design of product 10.1115/1.4046207.
families, Int. J. Prod. Res. 42 (2004) 2955–2969, https://doi.org/10.1080/ [80] F. Tao, J. Cheng, Q. Qi, M. Zhang, H. Zhang, F. Sui, Digital twin-driven product
00207540410001691929. design, manufacturing and service with big data, Int. J. Adv. Manuf. Technol. 949
[54] E.M. Grames, A.N. Stillman, M.W. Tingley, C.S. Elphick, An automated approach (94) (2018) 3563–3576, https://doi.org/10.1007/S00170-017-0233-1.
to identifying search terms for systematic reviews using keyword co-occurrence [81] Z. Wang, C.-H. Chen, P. Zheng, X. Li, L.P. Khoo, A graph-based context-aware
networks, Methods Ecol. Evol. 10 (2019) 1645–1654, https://doi.org/10.1111/ requirement elicitation approach in smart product-service systems, Int. J. Prod.
2041-210X.13268. Res. 59 (2019) 635–651, https://doi.org/10.1080/00207543.2019.1702227.
[55] S. Rose, D. Engel, N. Cramer, W. Cowley, Automatic Keyword Extraction from [82] Z. Wang, C.H. Chen, P. Zheng, X. Li, L.P. Khoo, A novel data-driven graph-based
Individual Documents, Text Min, Appl. Theory. (2010) 1–20, https://doi.org/ requirement elicitation framework in the smart product-service system context,
10.1002/9780470689646.CH1. Adv. Eng. Informatics. 42 (2019), 100983, https://doi.org/10.1016/J.
[56] D. Moher, A. Liberati, J. Tetzlaff, D.G. Altman, T.P. Group, Preferred Reporting AEI.2019.100983.
Items for Systematic Reviews and Meta-Analyses: The PRISMA Statement, PLOS [83] P. Zheng, Y. Liu, F. Tao, Z. Wang, C.H. Chen, Smart Product-Service Systems
Med. 6 (2009) e1000097. https://doi.org/10.1371/JOURNAL.PMED.1000097. Solution Design via Hybrid Crowd Sensing Approach, IEEE Access 7 (2019)
[57] M.J. Page, J.E. McKenzie, P.M. Bossuyt, I. Boutron, T.C. Hoffmann, C.D. Mulrow, 128463–128473, https://doi.org/10.1109/ACCESS.2019.2939828.
L. Shamseer, J.M. Tetzlaff, E.A. Akl, S.E. Brennan, R. Chou, J. Glanville, J.

16
D. Vlah et al. Advanced Engineering Informatics 54 (2022) 101774

[84] B. Song, B. Yan, G. Triulzi, J. Alstott, J. Luo, Overlay technology space map for [113] C. Lee, B. Song, Y. Park, Design of convergent product concepts based on
analyzing design knowledge base of a technology domain: the case of hybrid functionality: An association rule mining and decision tree approach, Expert Syst.
electric vehicles, Res. Eng. Des. 30 (2019) 405–423, https://doi.org/10.1007/ Appl. 39 (2012) 9534–9542, https://doi.org/10.1016/J.ESWA.2012.02.099.
S00163-019-00312-W. [114] F. Tao, J. Cheng, Q. Qi, M. Zhang, H. Zhang, F. Sui, Digital twin-driven product
[85] S. Sarica, B. Song, E. Low, J. Luo, Engineering Knowledge Graph for Keyword design, manufacturing and service with big data, Int. J. Adv. Manuf. Technol. 949
Discovery in Patent Search, Proc. Des. Soc. Int. Conf. Eng. Des. 1 (2019) (94) (2017) 3563–3576, https://doi.org/10.1007/S00170-017-0233-1.
2249–2258, https://doi.org/10.1017/DSI.2019.231. [115] S.R. John, C.S. Tucker, Quantifying the price and demand of subassemblies in the
[86] B. Song, J. Luo, K. Wood, Data-Driven Platform Design: Patent Data and Function end of life strategy of product resynthesis, Proc. ASME Des. Eng. Tech. Conf. 4
Network Analysis, J. Mech. Des. 141 (2019), https://doi.org/10.1115/ (2014), https://doi.org/10.1115/DETC2014-34757.
1.4042083. [116] K. Afrin, B. Nepal, L. Monplaisir, A data-driven framework to new product
[87] S. Sarica, J. Luo, K.L. Wood, TechNet: Technology Semantic Network Based on demand prediction: Integrating product differentiation and transfer learning
Patent Data, Expert Syst. Appl. 142 (2019), https://doi.org/10.1016/j. approach, Expert Syst. Appl. 108 (2018) 246–257, https://doi.org/10.1016/J.
eswa.2019.112995. ESWA.2018.04.032.
[88] Z. Li, Z.G. Tian, J.W. Wang, W.M. Wang, G.Q. Huang, Dynamic mapping of design [117] E.L.B. Christian, X. Zheng, S.R. Miller, Linking creativity measurements to
elements and affective responses: a machine learning based method for affective product market favorability: A data-mining approach, Proc. ASME Des. Eng. Tech.
design, J. Eng. Des. 29 (2018) 358–380, https://doi.org/10.1080/ Conf. 2A–2017 (2017), https://doi.org/10.1115/DETC2017-67622.
09544828.2018.1471671. [118] C. Obieke, J. Milisavljevic-Syed, J. Han, Supporting Design Problem-exploring
[89] J. Wang, W.M. Wang, Z. Tian, Z. Li, Classification of multiple affective attributes with Emergent Technologies, Procedia CIRP. 91 (2020) 373–381, https://doi.org/
of customer reviews: Using classical machine learning and deep learning, ACM 10.1016/J.PROCIR.2020.02.189.
Int. Conf. Proceeding Ser. (2018), https://doi.org/10.1145/3207677.3277953. [119] J. Luo, B. Song, L. Blessing, K. Wood, Design opportunity conception using the
[90] P. Kestel, P. Kügler, C. Zirngibl, B. Schleich, S. Wartzack, Ontology-based total technology space map, AI EDAM. 32 (2018) 449–461, https://doi.org/
approach for the provision of simulation knowledge acquired by Data and Text 10.1017/S0890060418000094.
Mining processes, Adv. Eng. Informatics. 39 (2019) 292–305, https://doi.org/ [120] D. Chang, C. Lee, A product affective properties identification approach based on
10.1016/J.AEI.2019.02.001. web mining in a crowdsourcing environment, Https://Doi.Org/10.1080/
[91] F. Dworschak, P. Kügler, B. Schleich, S. Wartzack, Integrating the Mechanical 09544828.2018.1463514. 29 (2018) 449–483. https://doi.org/10.1080/
Domain into Seed Approach, Proc. Des. Soc. Int. Conf. Eng. Des. 1 (2019) 09544828.2018.1463514.
2587–2596, https://doi.org/10.1017/DSI.2019.265. [121] Q. Sun, J. Niu, Z. Yao, H. Yan, Exploring eWOM in online customer reviews:
[92] C. Sauer, B. Schleich, S. Wartzack, Deep learning in sheet-bulk metal forming part Sentiment analysis at a fine-grained level, Eng. Appl. Artif. Intell. 81 (2019)
design, Proc. Int. Des. Conf. Des. 6 (2018) 2999–3010. https://doi.org/10. 68–78, https://doi.org/10.1016/J.ENGAPPAI.2019.02.004.
21278/IDC.2018.0147. [122] T. Jiang, Q. Duan, J. Zhu, H. Liu, L. Yu, Starch-based Biodegradable Materials:
[93] A. Wolf, N. Binder, J. Miehling, S. Wartzack, Towards Virtual Assessment of Challenges and Opportunities, Adv. Ind. Eng. Polym. Res. (2019), https://doi.
Human Factors: A Concept for Data Driven Prediction and Analysis of Physical org/10.1016/J.AIEPR.2019.11.003.
User-product Interactions, Proc. Des. Soc. Int. Conf. Eng. Des. 1 (2019) [123] W.M. Wang, J.W. Wang, Z. Li, Z.G. Tian, E. Tsui, Multiple affective attribute
4029–4038, https://doi.org/10.1017/DSI.2019.410. classification of online customer product reviews: A heuristic deep learning
[94] H. Wang, C. Tucker, Pixel to stroke sketch generation using reinforcement method for supporting Kansei engineering, Eng. Appl. Artif. Intell. 85 (2019)
learning, Proc. ASME Des. Eng. Tech. Conf. 1 (2019), https://doi.org/10.1115/ 33–45, https://doi.org/10.1016/J.ENGAPPAI.2019.05.015.
DETC2019-98481. [124] W.M. Wang, Z. Li, Z.G. Tian, J.W. Wang, M.N. Cheng, Extracting and
[95] S. Tuarob, C.S. Tucker, Quantifying Product Favorability and Extracting Notable summarizing affective features and responses from online product descriptions
Product Features Using Large Scale Social Media Data, J. Comput. Inf. Sci. Eng. and reviews: A Kansei text mining approach, Eng. Appl. Artif. Intell. 73 (2018)
15 (2015), https://doi.org/10.1115/1.4029562. 149–162, https://doi.org/10.1016/j.engappai.2018.05.005.
[96] C.S. Tucker, H.M. Kim, Data-driven decision tree classification for product [125] W. Kim, T. Ko, I. Rhiu, M.H. Yun, Mining affective experience for a kansei design
portfolio design optimization, J. Comput. Inf. Sci. Eng. 9 (2009) 1–14, https:// study on a recliner, Appl. Ergon. 74 (2019) 145–153, https://doi.org/10.1016/J.
doi.org/10.1115/1.3243634. APERGO.2018.08.014.
[98] C.E. Lopez, Z.V. Zhao, C.S. Tucker, Semantic Network Differences Across [126] W. han Choi, J. moon Kim, G.J. Park, Comparison study of some commercial
Engineering Design Communication Methods, Proc. ASME Des. Eng. Tech. Conf. 1 structural optimization software systems, Struct. Multidiscip. Optim. 54 (2016)
(2019), https://doi.org/10.1115/DETC2019-97728. 685–699. https://doi.org/10.1007/s00158-016-1429-y.
[99] A. Kusiak, Smart manufacturing, Int. J. Prod. Res. 56 (2018) 508–517, https:// [127] D. Raghupathi, B. Yannou, R. Farel, E. Poirson, Customer sentiment appraisal
doi.org/10.1080/00207543.2017.1351644. from user-generated product reviews: a domain independent heuristic algorithm,
[100] X. Wang, C.X. Feng, Development of Empirical Models for Surface Roughness Int. J. Interact. Des. Manuf. 93 (9) (2015) 201–211, https://doi.org/10.1007/
Prediction in Finish Turning, Int. J. Adv. Manuf. Technol. (2002 205(20), 2002,) S12008-015-0273-4.
348–356, https://doi.org/10.1007/S001700200162. [128] Y. Han, M. Moghaddam, Analysis of sentiment expressions for user-centered
[101] Q. Qi, F. Tao, Digital Twin and Big Data Towards Smart Manufacturing and design, Expert Syst. Appl. 171 (2021), 114604, https://doi.org/10.1016/J.
Industry 4.0: 360 Degree Comparison, IEEE, Access. 6 (2018) 3585–3593, https:// ESWA.2021.114604.
doi.org/10.1109/ACCESS.2018.2793265. [129] L. Zhang, X. Chu, D. Xue, Identification of the to-be-improved product features
[102] J. Zhan, L. Tong, Y. Liu, Gather customer concerns from online product reviews - based on online reviews for product redesign, Int. J. Prod. Res. 57 (2018)
A text summarization approach, Expert Syst. with Appl. An, Int. J. 36 (2009) 2464–2479, https://doi.org/10.1080/00207543.2018.1521019.
2107–2115, https://doi.org/10.1016/J.ESWA.2007.12.039. [130] N. El Dehaibi, N.D. Goodman, E.F. MacDonald, Extracting Customer Perceptions
[103] S. Ahmed, S. Kim, K.M. Wallace, A Methodology for Creating Ontologies for of Product Sustainability From Online Reviews, J. Mech. Des. 141 (2019), https://
Engineering Design, J. Comput. Inf. Sci. Eng. 7 (2007) 132–140, https://doi.org/ doi.org/10.1115/1.4044522.
10.1115/1.2720879. [131] R. Rai, Identifying key product attributes and their importance levels from online
[104] M.C. Pasqual, O.L. de Weck, Multilayer network model for analysis and customer reviews, Proc. ASME Des. Eng. Tech. Conf. 3 (2012) 533–540, https://
management of change propagation, Res. Eng. Des. 234 (23) (2011) 305–328, doi.org/10.1115/DETC2012-70493.
https://doi.org/10.1007/S00163-011-0125-6. [132] C. Hoyle, W. Chen, N. Wang, G. Gomez-Levi, Understanding and modelling
[105] D. Wu, J.L. Thames, D.W. Rosen, D. Schaefer, Enhancing the Product Realization heterogeneity of human preferences for engineering design, J. Eng. Des. 22
Process With Cloud-Based Design and Manufacturing Systems, J. Comput. Inf. Sci. (2010) 583–601, https://doi.org/10.1080/09544821003604496.
Eng. 13 (2013), https://doi.org/10.1115/1.4025257. [133] J. Zhang, A. Simeone, P. Gu, B. Hong, Product features characterization and
[106] K.C. Wang, A hybrid Kansei engineering design expert system based on grey customers’ preferences prediction based on purchasing data, CIRP Ann. 67 (2018)
system theory and support vector regression, Expert Syst. Appl. 38 (2011) 149–152, https://doi.org/10.1016/J.CIRP.2018.04.020.
8738–8750, https://doi.org/10.1016/J.ESWA.2011.01.083. [134] T. Hou, B. Yannou, Y. Leroy, E. Poirson, Mining customer product reviews for
[107] A. Kusiak, Innovation: The Living Laboratory Perspective, Comput. Aided. Des. product development: A summarization process, Expert Syst. Appl. 132 (2019)
Appl. 4 (2007) 863–876, https://doi.org/10.1080/16864360.2007.10738518. 141–150, https://doi.org/10.1016/J.ESWA.2019.04.069.
[108] F. Darses, M. Wolff, How do designers represent to themselves the users’ needs? [135] H.Q. Chen, T. Honda, M.C. Yang, Approaches for identifying consumer
Appl. Ergon. 37 (2006) 757–764, https://doi.org/10.1016/J. preferences for the design of technology products: A case study of residential solar
APERGO.2005.11.004. panels, J. Mech. Des. Trans. ASME. 135 (2013), https://doi.org/10.1115/
[109] S. Ravikumar, K.I. Ramachandran, V. Sugumaran, Machine learning approach for 1.4024232.
automated visual inspection of machine components, Expert Syst. Appl. 38 (2011) [136] M. Ozer, U. Cebeci, Affective design using big data within the context of online
3260–3266, https://doi.org/10.1016/J.ESWA.2010.09.012. shopping, J. Eng. Des. 30 (2019) 368–384, https://doi.org/10.1080/
[110] H.C. Chang, H.H. Lai, Y.M. Chang, A measurement scale for evaluating the 09544828.2019.1656803.
attractiveness of a passenger car form aimed at young consumers, Int. J. Ind. [137] M.B. Canciglieri, A.F.C.S. de Moura Leite, A.L. Szejka, O.C. Junior, An Approach
Ergon. 37 (2007) 21–30, https://doi.org/10.1016/J.ERGON.2006.09.014. to Semantic Interoperability for Product Development Through Automatic
[111] C.C. Aguwa, L. Monplaisir, O. Turgut, Voice of the customer: Customer Requirement Extraction and Semantic Reconciliation, IFIP Adv. Inf. Commun.
satisfaction ratio based analysis, Expert Syst. Appl. 39 (2012) 10112–10119, Technol. 565 IFIP (2019) 126–135. https://doi.org/10.1007/978-3-030-42250-9_
https://doi.org/10.1016/J.ESWA.2012.02.071. 12.
[112] J.K. Bae, J. Kim, Product development with data mining techniques: A case on [138] Y. Jiao, Y. Yang, J. Zhong, H. Zhang, A comparative analysis of intelligent
design of digital camera, Expert Syst. Appl. 38 (2011) 9274–9280, https://doi. classifiers for mapping customer requirements to product configurations, ACM
org/10.1016/J.ESWA.2011.01.030.

17
D. Vlah et al. Advanced Engineering Informatics 54 (2022) 101774

Int. Conf. Proceeding Ser. Part F 1325 (2017) 72–77, https://doi.org/10.1145/ [164] C. He, Z. kai Li, S. Wang, D. zhuo Liu, A systematic data-mining-based
3152723.3152726. methodology for product family design and product configuration, Adv. Eng.
[139] A.M. Bigorra, O. Isaksson, Combining customer needs and the customer’s way of Informatics. 48 (2021) 101302. https://doi.org/10.1016/J.AEI.2021.101302.
using the product to set customer-focused targets in the House of Quality, Int. J. [165] B. Agard, A. Kusiak, Data-mining-based methodology for the design of product
Prod. Res. 55 (2016) 2320–2335, https://doi.org/10.1080/ families, Int. J. Prod. Res. 42 (2007) 2955–2969, https://doi.org/10.1080/
00207543.2016.1238114. 00207540410001691929.
[140] J.R. Li, Q.H. Wang, A rough set based data mining approach for house of quality [166] S.K. Moon, S.R.T. Kumara, T.W. Simpson, Data mining and fuzzy clustering to
analysis, Int. J. Prod. Res. 48 (2010) 2095–2107, https://doi.org/10.1080/ support product family design, Proc. ASME Des. Eng. Tech. Conf. 2006 (2006),
00207540802665907. https://doi.org/10.1115/DETC2006-99287.
[141] V.J. Ferrero, N. Alqseer, M. Tensa, B. DuPont, Using Decision Trees Supported by [167] T. Kipp, D. Krause, Computer Aided Size Range Development - Data Mining vs.
Data Mining to Improve Function-Based Design, Proc. ASME Des. Eng. Tech. Conf. Optimization, DS 58-4 Proc. ICED 09, 17th Int. Conf. Eng. Des. Vol. 4, Prod. Syst.
11A–2020 (2020), https://doi.org/10.1115/DETC2020-22498. Des. Palo Alto, CA, USA, 24.-27.08.2009. (2009) 179–190.
[142] L.L. Zhang, Identifying mapping relationships between functions and technologies [168] J. Hochdörffer, C. Laule, G. Lanza, Product variety management using data-
with association rule mining, Https://Doi.Org/10.1080/0951192X.2011.637962. mining methods - Reducing planning complexity by applying clustering analysis
25 (2012) 496–508. https://doi.org/10.1080/0951192X.2011.637962. on product portfolios, IEEE Int. Conf. Ind. Eng. Eng. Manag. (2017-December
[143] M. Tensa, K. Edmonds, V. Ferrero, A. Mikes, N. Soria Zurita, R. Stone, B. DuPont, (2018)) 593–597, https://doi.org/10.1109/IEEM.2017.8289960.
Toward automated functional modeling: An association rules approach for mining [169] C.J. Romanowski, R. Nagi, A Data Mining Approach to Forming Generic Bills of
the relationship between product components and function, Proc. Int. Conf. Eng. Materials in Support of Variant Design Activities, J. Comput. Inf. Sci. Eng. 4
Des. ICED. (2019-August (2019)) 1713–1722, https://doi.org/10.1017/ (2004) 316–328, https://doi.org/10.1115/1.1812556.
DSI.2019.177. [170] M. Schmidt, B. Gehring, J.-S. Gerber, J.M. Stocker, M. Kreimayer, M. Lienkamp,
[144] S.W. Kang, C. Tucker, An automated approach to quantifying functional Graph-based similarity analysis of BOM data to identify unnecessary inner
interactions by mining large-scale product specification data, Http://Dx.Doi.Org/ product variance / The Design Society, in: Proc. 21st Int. Conf. Eng. Des. (ICED
10.1080/09544828.2015.1083539. 27 (2015) 1–24, https://doi.org/10.1080/ 17), 2017.
09544828.2015.1083539. [171] Y.M. Zhao, D.S. Liu, Z.J. Wen, Optimization design method of product general
[145] Z. Zhang, L. Liu, W. Wei, F. Tao, T. Li, A. Liu, A Systematic Function tolerance system, Int. J. Adv. Manuf. Technol. 701 (70) (2013) 363–374, https://
Recommendation Process for Data-Driven Product and Service Design, J. Mech. doi.org/10.1007/S00170-013-5193-5.
Des. Trans. ASME. 139 (2017), https://doi.org/10.1115/1.4037610. [172] Y. Yu, D. Hu, H. Wang, G. Zhao, Data mining and recommendation of engineering
[146] J. Han, D. Park, H. Forbes, D. Schaefer, A computational approach for using social note items in MBD dataset, ACM Int. Conf. Proceeding Ser. (2019) 1–6, https://
networking platforms to support creative idea generation, Procedia CIRP. 91 doi.org/10.1145/3313991.3314015.
(2020) 382–387, https://doi.org/10.1016/J.PROCIR.2020.02.190. [173] R. Kretschmer, A. Pfouga, S. Rulhoff, J. Stjepandić, Knowledge-based design for
[147] D. Chang, C.H. Chen, Product concept evaluation and selection using data mining assembly in agile manufacturing by using Data Mining methods, Adv. Eng.
and domain ontology in a crowdsourcing environment, Adv. Eng. Informatics. 29 Informatics. 33 (2017) 285–299, https://doi.org/10.1016/J.AEI.2016.12.006.
(2015) 759–774, https://doi.org/10.1016/J.AEI.2015.06.003. [174] O. Abdelrahman, P. Keikhosrokiani, Assembly line anomaly detection and root
[148] B. Camburn, Y. He, S. Raviselvam, J. Luo, K. Wood, Evaluating Crowdsourced cause analysis using machine learning, IEEE Access 8 (2020) 189661–189672,
Design Concepts With Machine Learning, Proc. ASME Des. Eng. Tech. Conf. 7 https://doi.org/10.1109/ACCESS.2020.3029826.
(2019), https://doi.org/10.1115/DETC2019-97285. [175] B. Dolunay, A. Akgunduz, Automated end-user behaviour assessment tool for
[149] L. Siddharth, N. Madhusudanan, A. Chakrabarti, Toward Automatically Assessing remote product and system testing, Expert Syst. Appl. 34 (2008) 2511–2523,
the Novelty of Engineering Design Solutions, J. Comput. Inf. Sci. Eng. 20 (2020), https://doi.org/10.1016/J.ESWA.2007.04.011.
https://doi.org/10.1115/1.4044318. [176] A. Lugmayr, S. Bender, Free UX testing tool: The LudoVico UX Machine for
[150] M.E. Lynch, S. Sarkar, K. Maute, Machine Learning to Aid Tuning of Numerical physiological sensor data recording, analysis, and visualization for user
Parameters in Topology Optimization, J. Mech. Des. 141 (2019), https://doi.org/ experience design experiments, Conf. Hum. Factors Comput. Syst. - Proc. (07-
10.1115/1.4044228. May-2016 (2016)) 36–41, https://doi.org/10.1145/2898365.2899801.
[151] D. Nagaraj, D. Werth, Towards a Generalized System for Generative Engineering, [177] I.Ö. Arnarsson, O. Frost, E. Gustavsson, D. Stenholm, M. Jirstrand, J. Malmqvist,
ACM Int. Conf. Proceeding Ser. (2020), https://doi.org/10.1145/ Supporting Knowledge Re-Use with Effective Searches of Related Engineering
3378184.3378227. Documents - A Comparison of Search Engine and Natural Language Processing-
[152] L. Graening, B. Sendhoff, Shape mining: A holistic data mining approach for Based Algorithms, Proc. Des. Soc. Int. Conf. Eng. Des. 1 (2019) 2597–2606,
engineering design, Adv. Eng. Informatics. 28 (2014) 166–185, https://doi.org/ https://doi.org/10.1017/DSI.2019.266.
10.1016/J.AEI.2014.03.002. [178] P. Zheng, C.H. Chen, S. Shang, Towards an automatic engineering change
[153] Z. Zhang, L. Gong, Y. Jin, J. Xie, J. Hao, A quantitative approach to design management in smart product-service systems – A DSM-based learning approach,
alternative evaluation based on data-driven performance prediction, Adv. Eng. Adv. Eng. Informatics. 39 (2019) 203–213, https://doi.org/10.1016/J.
Informatics. 32 (2017) 52–65, https://doi.org/10.1016/J.AEI.2016.12.009. AEI.2019.01.002.
[154] S.K. Dasari, A. Cheddad, P. Andersson, Random Forest Surrogate Models to [179] J. Milisavljevic-Syed, J. Lane Thames, D. Schaefer, The Digitization of Design and
Support Design Space Exploration in Aerospace Use-Case, IFIP Adv. Inf. Commun. Manufacturing: A State-of-the-Art Report on the Transition from Strategic Vision
Technol. 559 (2019) 532–544, https://doi.org/10.1007/978-3-030-19823-7_45. to Implementation in Industry, Procedia CIRP. 93 (2020) 575–580, https://doi.
[155] A. Angrish, A. Bharadwaj, B. Starly, MVCNN++: Computer-Aided Design Model org/10.1016/J.PROCIR.2020.03.088.
Shape Classification and Retrieval Using Multi-View Convolutional Neural [180] G. Schuh, C. Kelzenberg, J. Wiese, Design model for the cost calculation of
Networks, J. Comput. Inf. Sci. Eng. 21 (2021), https://doi.org/10.1115/ product-service systems in single and small series production, Procedia CIRP. 84
1.4047486. (2019) 296–301, https://doi.org/10.1016/J.PROCIR.2019.04.216.
[156] R. Ben Hadj, I. Belhadj, C. Gouta, M. Trigui, N. Aifaoui, M. Hammadi, An [181] V.N. Andreev, M.A. Charuyskaya, A.S. Kryzhanovskaya, I.D. Mursalov, A.
interoperability process between CAD system and CAE applications based on CAD A. Mozharovskaia, S.G. Chervenkova, Application of intelligent engineering in the
data, Int. J. Interact. Des. Manuf. 2017 123. 12 (2017) 1039–1058. https://doi. planning of cyber-physical production systems, Int. J. Adv. Manuf. Technol. 115
org/10.1007/S12008-017-0445-5. (2021) 117–123, https://doi.org/10.1007/S00170-021-07108-0.
[157] A. Kreis, M. Hirz, CAD-Automation in Automotive Development – Potentials, [182] Z. Zhu, K. Ferreira, N. Anwer, L. Mathieu, K. Guo, L. Qiao, Convolutional Neural
Limits and Challenges, Comput. Aided Des. Appl. 18 (2021) 56–60. https://doi. Network for geometric deviation prediction in Additive Manufacturing, Procedia
org/10.14733/CADCONFP.2020.56-60. CIRP. 91 (2020) 534–539, https://doi.org/10.1016/J.PROCIR.2020.03.108.
[158] S. Bickel, C. Sauer, B. Schleich, S. Wartzack, Comparing CAD part models for [183] G.J. Duan, X. Yan, A Real-Time Quality Control System Based on Manufacturing
geometrical similarity: A concept using machine learning algorithms, Procedia Process Data, IEEE Access 8 (2020) 208506–208517, https://doi.org/10.1109/
CIRP. 96 (2020) 133–138, https://doi.org/10.1016/J.PROCIR.2021.01.065. ACCESS.2020.3038394.
[159] G. Vasantha, D. Purves, J. Quigley, J. Corney, A. Sherlock, G. Randika, Common [184] A. Balu, S. Ghadai, S. Sarkar, A. Krishnamurthy, Orthogonal distance fields
design structures and substitutable feature discovery in CAD databases, Adv. Eng. representation for machine-learning based manufacturability analysis, Proc.
Informatics. 48 (2021), 101261, https://doi.org/10.1016/J.AEI.2021.101261. ASME Des. Eng. Tech. Conf. 9 (2020), https://doi.org/10.1115/DETC2020-
[160] Y. Ishino, Y. Jin, An information value based approach to design procedure 22487.
capture, Adv. Eng. Informatics. 20 (2006) 89–107, https://doi.org/10.1016/J. [185] Y. Fathy, M. Jaber, A. Brintrup, Learning with Imbalanced Data in Smart
AEI.2005.04.002. Manufacturing: A Comparative Analysis, IEEE Access 9 (2021) 2734–2757,
[161] T. Spruegel, R. Rothfelder, S. Bickel, A. Grauf, C. Sauer, B. Schleich, S. Wartzack, https://doi.org/10.1109/ACCESS.2020.3047838.
Methodology for plausibility checking of structural mechanics simulations using [186] L. Ruiz, M. Torres, A. Gómez, S. Díaz, J.M. González, F. Cavas, Detection and
deep learning on existing simulation data, in: Proc. Nord., Linköping, Sweden, Classification of Aircraft Fixation Elements during Manufacturing Processes Using
2018. a Convolutional Neural Network, Appl. Sci. 2020, Vol. 10, Page 6856. 10 (2020)
[162] Y. Shao, H. Zhu, R. Wang, Y. Liu, Y. Liu, A Simulation Data-Driven Design 6856. https://doi.org/10.3390/APP10196856.
Approach for Rapid Product Optimization, J. Comput. Inf. Sci. Eng. 20 (2020), [187] W. Wei, J. Yuan, A. Liu, Manufacturing data-driven process adaptive design
https://doi.org/10.1115/1.4045527. method, Procedia CIRP. 91 (2020) 728–734, https://doi.org/10.1016/J.
[163] X. Shi, J. Chen, H. Yang, Y. Peng, X. Ruan, A novel approach to extract knowledge PROCIR.2020.02.230.
from simulation results, Int. J. Adv. Manuf. Technol. 20 (2002) 390–396, https:// [188] D. Ikeuchi, A. Vargas-Uscategui, X. Wu, P.C. King, Data-Efficient Neural Network
doi.org/10.1007/S001700200168. for Track Profile Modelling in Cold Spray Additive Manufacturing, Appl. Sci.
2021, Vol. 11, Page 1654. 11 (2021) 1654. https://doi.org/10.3390/
APP11041654.

18
D. Vlah et al. Advanced Engineering Informatics 54 (2022) 101774

[189] L. Romeo, J. Loncarski, M. Paolanti, G. Bocchini, A. Mancini, E. Frontoni, implementation, Proc. Int. Des. Conf. Des. 4 (2018) 1571–1582, https://doi.org/
Machine learning-based design support system for the prediction of 10.21278/IDC.2018.0399.
heterogeneous machine parameters in industry 4.0, Expert Syst. Appl. 140 [192] J. Nelson, S. Buisine, A. Aoussat, Anticipating the use of future things: Towards a
(2020), https://doi.org/10.1016/J.ESWA.2019.112869. framework for prospective use analysis in innovation design projects, Appl.
[190] J. Lützenberger, P. Klein, K. Hribernik, K.D. Thoben, Improving Product-Service Ergon. 44 (2013) 948–956, https://doi.org/10.1016/J.APERGO.2013.01.002.
Systems by Exploiting Information from the Usage Phase, A Case Study, Procedia [193] L. He, C. Hoyle, W. Chen, J. Wang, B. Yannou, A framework for choice modeling
CIRP. 47 (2016) 376–381, https://doi.org/10.1016/J.PROCIR.2016.03.064. in usage context-based design, Proc. ASME Des. Eng. Tech. Conf. 1 (2010)
[191] C. Hollauer, B. Shalumov, J. Wilberg, M. Omer, Graph databases for exploiting 265–276, https://doi.org/10.1115/DETC2010-28490.
use phase data in product-service-system development: A methodology to support

19