A Feasibility Discussion: Is ML Suitable for Predicting Sustainable Patterns in Consumer Product Preferences?

Abstract

In the era when product design must meet the needs of consumers, the products preferred by consumers are an important source of design creativity and design reference for product designers to design products. Therefore, how to effectively grasp the products that consumers prefer has become an important issue for product designers. In order to allow designers to have more convenient and accurate consumer preference product prediction tools, this study proposed machine learning (ML) to analyze and predict sustainable patterns in consumer product preferences and conducted a feasibility study on the use of ML for predicting sustainable patterns in consumer product preferences. A total of three experiments were carried out in this study: the KJ method to predict consumer product preference experiment, the AHP method to predict consumer product preference experiment, and ML to predict consumer product preference experiment. This study uses the three experiments to discuss and compare the prediction ability of ML and the current commonly used forecasting tools, namely the KJ method and AHP method. The research results show that no matter what kind of consumer product attribute preference is predicted, the accuracy rate of consumer product preference prediction by ML is much higher than that of the KJ method and AHP method. These research results show that no matter the product attribute dimension, ML has the ability to predict consumer preferences, and ML has a better ability to predict consumer preferences than traditional tools. Therefore, this study believes that ML can be used to analyze and predict sustainable patterns in consumer product preferences. Therefore, this study suggests that product designers can use ML technology to assist in the analysis and prediction of consumer product preferences, so as to improve the grasp of consumer preference products.