Topic Mining Based on Online Shopping Users' Reviews

Digging into the potential features of user comments is of great significance to optimize product design and enhance market competitiveness. Taking baby pacifiers as an example, this paper uses the user review data on Amazon platform to extract the topic features of the user review content by constructing LDA model, and measures the intensity of related topics. The results show that user comments mainly include six topics, such as product quality, color, delivery time, appearance, service and price. Among them, the product quality and price topic intensity are the highest, the product color, delivery time and service topic intensity are relatively high, and the product appearance topic intensity is the lowest. Abstracting and quantifying the topic features helps enterprises understand the purchase needs of online shoppers, and then optimize product design and make reasonable strategic adjustments to promote their own long-term development.