Research On Online Consumer Preference Based On Emoji And Text Mining

The rich commodities on the online shopping platform provide greater choices for online consumers,but at the same time it also brings about the problem of over-selection.To solve the problems caused by information overload,enterprises often provide personalized recommendation services based on the preferences of online consumers.How to accurately understand and mine online consumer preferences is the key to personalized recommendations.Online reviews are direct feedback from online consumers after purchasing goods,and contain rich preference information,which can be used as an effective data source for preference mining.However,traditional research often only considers the sentiment of the comment text and ignores the role of emoji.Therefore,this article will integrate emoji and online comment text to mine online consumer preferences.First of all,it combs the domestic and foreign literatures on Internet consumer preference research,proposes online consumer preference mining methods based on online review texts,constructs sentiment dictionaries,degree word dictionaries and negative word dictionaries to mine online consumers' different attributes of commodities Preferences.Then,it proposes a method for mining online consumer preferences that integrates emoticons and online comment texts,constructs an emoticon sentiment dictionary to obtain expression emotions,and integrates emoticons and text to establish a fine-grained attribute sentiment analysis method to mine online consumers' attribute preferences.Finally,using the actual online review data of public comment hotels in Guilin to conduct a case analysis,the research results show that: through online review mining of 20 hotels in Guilin,online consumers' preference for hotel attributes is mainly reflected in location,services,facilities,Seven aspects of catering,environment,transportation and price;integrated emoji and online comment text preference mining method can effectively mine online consumer preferences,and the accuracy rate is 11% higher than that based on online comment text preference mining method.Integrating emoticons and online review texts to deeply dig network consumers' preferences for product attributes,not only helps to understand the degree of online consumers' preference for different aspects of commodities,but also helps companies analyze the online consumers' demand and preference laws To provide a certain theoretical basis for online marketing and personalized recommendation services.