Data Mining of New Snack E-commerce Reviews Based on Text Sentiment Analysis and Latent Dirichlet Allocation Topic Model
Yang Qian · 2020
In the new retail era, in order to promote the development of new snack e-commerce enterprises, increase customer satisfaction with their products and services and the desire to repurchase, this article makes sentiment analysis on the 430,000 online reviews of new snack e-commerce as crawled online from the perspective of customers.Firstly, it makes an overall sentiment judgment on the online review texts of five stores.Secondly, it constructs the Latent Dirichlet Allocation (LDA) topic model, gets ten categories of keywords, analyzes the cluster analysis results and discusses for improving any consumer dissatisfaction point.Empirical analysis shows that in order to improve customer satisfaction and stickiness and attract more potential consumers, new snack e-commerce companies need to pay attention to optimizing brand reputation, product quality, service level, marketing strategy, and review information.