Research On Consumer Emotion Analysis And Influencing Factors Of Aquatic Product E-commerce Based On Text Mining

The popularity and development of the Internet has driven the rapid development and growth of e-commerce.The development scale of fresh e-commerce in China is also increasing.As an important part of fresh e-commerce,aquatic e-commerce has a rapid development momentum.Due to the limitations of time and space,aquatic products have high requirements for refrigeration conditions and freshness.The vigorous development of aquatic e-commerce has broken this limitation,brought development opportunities for businesses and e-commerce platforms,and also provided convenience for consumers to buy aquatic products.The development of online shopping platform not only meets the diversified consumption needs of consumers,but also provides a channel for consumers to exchange information on the online shopping platform.As a subjective feeling of consumers after purchasing goods,online comments provide information support for consumers to understand products and businesses’ after-sales service,and have a great impact on the purchase behavior and attitude of potential consumers.Through literature review,there are many studies on online comments on electronic products,hotels and online public opinion,while the research on online comments in the field of aquatic products lags behind.In addition,it is difficult for consumers to obtain valuable information from various and miscellaneous online comment information.They often can not get the desired information through ordinary manual reading.With the development of Internet and computer technology,text mining algorithm provides a solution to this kind of problem.Therefore,this paper takes aquatic products as the research object,and uses text mining technology to mine the potential information in network comments.This paper selects the comment data of seafood and aquatic products in Jingdong fresh food channel,and uses Python software to analyze its emotion and LDA theme model.Firstly,we use Python to write code,crawl the comments of seafood and aquatic products on JD platform as the data source,and preprocess the comment data,including comment cleaning,de duplication,word segmentation and de stop words;Secondly,the emotional analysis of the comment data mainly uses two methods: emotional dictionary and machine learning.The crawled data is divided into positive and negative categories by using the emotional dictionary matching method.In order to judge the accuracy of the classification of emotional dictionary and compare with the classification results of machine learning,some comment data are extracted for emotional labeling,Compared with the emotion labels obtained from the emotion dictionary,the accuracy of the emotion dictionary method is obtained;The machine learning method is mainly to build a support vector machine model and a naive Bayesian model,evaluate the classification results of the two methods through indicators such as recall rate,accuracy rate,accuracy and F1 value,and then compare the three methods to select a classifier with better classification effect for emotional classification,so as to judge consumers’ emotional tendency towards JD aquatic products;Then,the classified positive and negative comments are displayed in the word cloud diagram,so as to intuitively understand the potential information in the aquatic product comment data in the form of graphics;Finally,the positive and negative comment data are analyzed by LDA theme model to explore the main factors affecting consumers’ different emotional tendencies.These factors also reflect the main characteristics of Jingdong aquatic products and consumers’ concerns about aquatic products.From the perspective of affective analysis methods,the classification effects of the three affective analysis methods can be expressed as follows: support vector machine >naive Bayes > affective dictionary,which shows that support vector machine has the highest accuracy in text classification;From the perspective of consumers’ emotional tendency towards aquatic products on the Jingdong platform,70% of consumers have a positive emotional attitude towards aquatic products on the Jingdong platform,but 30%of consumers still have a negative emotional attitude;from the display of positive and negative comment words and LDA theme mining,it can be seen that the factors affecting consumers’ positive emotions are mainly the freshness of aquatic products,taste and taste,logistics and distribution speed,and the high cost performance of aquatic products on the Jingdong platform.The factors that affect consumers’ negative emotions mainly include the quality of aquatic products,the quality of after-sales service,packaging protection and price differences.Based on the conclusions of the study,this paper makes different recommendations for consumers,merchants and platforms.For consumers,one is that Jingdong is a good choice in the case of many fresh e-commerce platforms that cannot be decided,the second is to buy fresh aquatic products to understand the product damage compensation situation,the third is to pay more attention to the preferential activities of merchants,to buy at the right time;for the Jingdong platform,one is to continue to improve and optimize the construction of the logistics system,the second is to strengthen packaging management,the third is to improve the after-sales service system and improve the consumer experience;for businesses,one is to strictly control product quality,to ensure the freshness and freshness of products,The second is to predict the activities in advance,reasonably optimize the price,and improve their own services.