Abstract:

With the rapid development of the Internet, online product recommendation has become a significant way of marketing business, and therefore the research on consumer preferences is the important basis for implementing product recommendation 32 product attributes in 7 categories are selected from catering group-buying products and the consumer personalized preference model based on product attributes is established by means of BP （back propagation）neural network Through questionnaire survey, the empirical study results show that the model can accurately predict individual consumer’s preference for a product, and that the most important factors influencing consumers’ preferences are price and the time of arrival Finally, some marketing implications are provided for businesses to implement product recommendation better

Key words: product attributes, personalized preference, group-buying;back propagation neural network