Affective design using machine learning: a survey and its prospect of conjoining big data

Abstract

Customer satisfaction in purchasing new products is an important issue that needs to be addressed in today's competitive markets. Consumers not only need to be solely satisfied with the functional requirements of a product, and they are also concerned with the affective needs and aesthetic appreciation of the product. A product with good affective design excites consumer emotional feelings so as to buy the product. However, affective design often involves complex and multi-dimensional problems for modelling and maximising affective satisfaction of customers. Machine learning is commonly used to model and maximise the affective satisfaction, since it is effective in modelling nonlinear patterns when numerical data relevant to the patterns is available. This article presents a survey of commonly used machine learning approaches for affective design when two data streams namely traditional survey data and modern big data are used. A classification of machine learning technologies is first provided which is developed using traditional survey data for affective design. The limitations and advantages of each machine learning technology are also discussed and we summarize the uses of machine learning technologies for affective design. This review article is useful for those who use machine learning technologies for affective design. The limitations of using traditional survey data are then discussed which is time consuming to collect and cannot fully cover all the affective domains for product development. Nowadays, big data related to affective design can be captured from social media. The prospects and challenges in using big data are discussed so as to enhance affective design, in which very limited research has so far been attempted. This article provides guidelines for researchers who are interested in exploring big data and machine learning technologies for affective design.

Use and Adaptations of Machine Learning in Big Data—Applications in Real Cases in Agriculture

“…There are many challenges to be overcome in developing a new ML algorithm or adapting classic algorithms to the context of Big Data. There are several works that detail these challenges [15,21,[38][39][40].…”

“…Chan [38] presents a survey on ML approaches commonly used for affective design when using two data streams, traditional survey data and modern big data. The author provides a classification of ML technologies for traditional survey data.…”

Predictive Analytics Using Social Big Data and Machine Learning

“…The Gartner magic quadrant for data science and machine learning platforms indicate a graphical competitive positioning of four categories of distinguished technology vendors for data science. These categories or quadrants are: Challengers, Leaders, Niche Players and Visionaries 1 . As depicted in Figure 2 sixteen different technology providers have been divided into the designated four categories.…”

Semantic Data Discovery from Social Big Data

“…The currently proliferated KGs span to embody domain dependent and independent facts. Examples of domain-independent (openworld) KGs include Freebase 8 , YAGO 9 , Dublin Core (DC 10 ), Simple Knowledge Organization System (SKOS 11 ), Semantically-Interlinked Online Communities (SIOC 12 ), and DBPedia 13 knowledge base, to cite a few. On the other hand, domaindependent KGs provide an overabundance of benefits to tackle domain-specific problems as well as to acquire the added value from domain corpora [55].…”