A Study on Shape Imagery of Electric Kettles Based on Quantification Theory Type I and Back Propagation Neural Networks

Abstract

In recent years, the market for electric kettles in China has been gradually expanding. As their functional needs for electric kettles are met, consumers are increasingly interested in the perceptual imagery conveyed by the shape of products. In order to meet consumers' aesthetic and perceptual needs for the shape of electric kettles, Back propagation Neural Network (BPNN) and natural language processing (NLP) technologies are integrated into the process of Kansei engineering to fully explore the shape imagery of electric kettles that meets consumers' needs. Firstly, a total of 81,487 consumer reviews of 189 electric kettle samples were captured from JD.com using web crawler tools and related technologies. Secondly, three groups of representative perceptual words were extracted from the reviews using NLP, Word2vec, factor analysis, cluster analysis, etc; multidimensional scaling and cluster analysis were used to analyse the collected sample products and 30 of them were selected as representatives. Additionally, the semantic difference scale was used to effectively evaluate consumers' perceptual images of the electric kettles, and a total of 306 valid survey results were collected, and the relationship model between the shape elements of electric kettle and consumers' perceptual images was constructed through quantification theory type I(QTTI) and BPNN, and the effects of the predictive models was compared. The results showed that the QTTI model is more accurate. By combining neural networks and natural language processing technologies, this study offered objective and systematically analysis of the relationship between the shape elements of electric kettles and consumer perceptual cognition of these elements, providing clear design directions and innovative ideas for the designers of related products.