A personalized recommendation model for online apparel shopping based on Kansei engineering. Issue 1 (6th March 2017)

• Record Type: Journal Article
• Authors: Zhou, Xiaoxi; Liang, Hui'e; Dong, Zhiya
• Is Part Of: International journal of clothing science and technology. Volume 29:Issue 1(2017)
• Page Start: 2
• Page End: 13
• Publication Date: 2017-03-06
• Subjects: Kansei engineering -- Big data -- Collaborative filtering -- Implicit knowledge -- Recommendation model; Clothing and dress -- Periodicals; Textile fabrics -- Periodicals; 677
• DOI: 10.1108/IJCST-12-2015-0137
• Languages: English
• ISSNs: 0955-6222

Abstract

Abstract : Purpose: Today clothing has become the largest category in online shopping in China, and even in Asia-Pacific. The satisfaction degree of apparel online shopping can be improved by effective personalized recommendation. The purpose of this paper is to propose a personalized recommendation model and algorithm based on Kansei engineering, traditional filtering algorithm and the knowledge relating to apparel. Design/methodology/approach: Users' perceptual image and the design elements of apparel based on Kansei engineering are discussed to build the mapping relation between the design elements and user ratings employing verbal protocol, semantic differential and partial least squares. The implicit knowledge and emotional needs pertaining to users are accessed using analytic hierarchy process. A personalized recommendation model for apparel online shopping is established and the algorithm for the personalized recommendation process is proposed. To present the personalized recommendation model, men's plaid shirts are taken as the example, and the recommendations of apparel for online shopping were implemented and ranked in the context of differing users' emotional needs. A comparison between the traditional model and this model is made to verify the effectiveness. Findings: The recommendation model is capable of analyzing data and information effectively, and providing fast, personalized apparel recommendation services in accordance with users' emotional needs. The experimental results suggest that the model is effective. Originality/value: Similar researches of recommendation mainly focus on the field of computer science, the basic idea of which is using users' history accessing records or the preferences of other similar users for determination of users' preferences. Since the attributes of apparel products are not factored in the approach referred above, the issue of personalized recommendation cannot be solved in a really effective way. Combining Kansei engineering and recommendation algorithm, a framework for apparel product recommendation is presented and it is a new way for improvement of recommendations for apparel products on shopping sites.