Predicting Consumer Preferences by the Deformation Threshold of Product Appearance

Abstract

Product appearance is essential in shaping consumers’ preferences. Thus, we predicted consumer preference by establishing a feasible threshold for appearance innovation with a novel method. We used deep learning (DL)-based image classification and the Waikato Environment for Knowledge Analysis (WEKA) as the implementation tool. Advanced DL algorithms were adopted to classify images of disparate products according to their aesthetics. The results of the classification presented that the method could explain the level of innovation in appearance that consumers favored. The result can be used as a guideline for industrial designers in the creation of new products. Moreover, the method can be used to assess existing products that need enhancement to meet consumer preferences. The proposed method assists industrial designers in understanding consumer preferences and product innovation. Future research is required to ensure the successful design of new products.

Keywords

• industrial design
• product development
• deep learning
• image recognition
• consumer preferences