Analysis Of E-commerce Fresh Food Sales Data Based On User Review

In recent years,due to the rapid development and wide application of ecommerce network platforms,people’s consumption patterns have changed greatly,thus a lot of people’s shopping channels have migrated from physical stores to online stores.After users purchase products in the online store,they will generate a large number of product-related reviews based on their experience,which contain a lot of emotional information,whether it is in product comparison and recommendation or product quality and monitoring.Under this condition,mining user evaluation information on the Internet has also become a hot spot for research,which can not only help other consumers choose the right products more e ciently,but also motivate merchants to improve product quality and service level,thereby enhancing market competitiveness.This article selects the products with high sales quantity of crab data in the fresh products of Jingdong Mall.First of all,the octopus software was used to crawl it,and more than 10,000 pieces of data collected were cleaned,preprocessed and featured.What’s more,the cleaned data were classified by machine learning methods such as Naive Bayes classification model,support vector machine classification model and random forest classification model,and the divided training data and test data were trained and tested according to the above three model algorithms,according to the calculation results of statistical model evaluation indicators such as accuracy and recall were compared,based on the experimental results showed that the Naive Bayes classification model had the best effect on sentiment classification.Finally,in order to more intuitively obtain the components of positive and negative comments affecting users,the LDA theme model is used to extract the subject words of positive comments and negative comments respectively,which can not only effectively reduce the time cost of users browsing user review information,but also put forward reasonable improvement suggestions and opinions for businesses.