Abstract As existing empirical researches mostly use the methods of questionnaire survey and experimental analysis, we use the data of B2C e-commerce sites, extract and represent the product features from four dimensions of perceived value including quality perception value, price perception value, service perception value and social perception value, and use the deep learning model to construct the consumer preference prediction model as well as explore the predictive explanatory power of each dimension of perceived value to consumer preference. The result shows that the CNN consumer preference prediction model based on perceived value is superior to the linear regression, random forest regression and other baseline algorithm. In the B2C online shopping environment, price perceptive value has the highest predictive explanatory power for consumer preference when purchasing durable goods, followed by quality perceptive value. Most consumers first care about product price and product function. Service perceptive value has the lowest predictive explanatory power for consumer preference.
LI Weiqing,CHI Maomao,WANG Weijun. Research on Online Consumer Preference Prediction Based on Perceived Value[J]. Chinese Journal of Management, 2021, 18(6): 912-.
