Design And Implementation Of Reviews Sentiment Analysis Subsystem In An Enterprise Merchant System

Sentiment analysis system is one of the more popular applications in the realm of natural language processing recently,and it has been developing continuously.The development of mobile Internet has greatly improved everyone’s participation.For example,when shopping online in an electronic mall,many people will fill in a few simple sentences as comments after receiving the product;for example,after using online ordering in a restaurant,many people will also leave their comments on the online ordering software.On the one hand,these online review data allow each user to clearly see the quality of service provided by each merchant,and at the same time make a merchant aware of their own problems.The main issue to be solved in this topic is to allow a user or businessman to more easily see the emotions contained in the comment data when faced with a large amount of online comment data.Through the investigation of the industry sentiment analysis system,this paper conducts technology selection and demand analysis,selects the current mainstream development framework and expounds the reasons for choosing and using it,and then divides users into roles and makes a full range of requirements for the system.analyze.Then,the general design of the system is carried out,the structure,database and interface of the system are designed,as well as,the functional modules are partitioned,mainly including sentiment polarity analysis,sentiment tendency analysis,comment rating prediction,login registration,system management and personal center and other modules,as well as the design of Bayesian-based sentiment analysis algorithm.In the section of detailed design,the devise of each module is introduced in detail by means of sequence diagrams and class diagrams,and the realization ideas and effects are shown.Finally,the system test work is introduced in an all-round way,and the test results are analyzed.The online comment sentiment analysis system can effectively mine and analyze the sentiment polarity and sentiment tendency contained in the text input by the user,and display the analysis results to the user clearly and friendly,so as to solve the problems that the sentiment analysis system needs to deal with.Through the design and implementation of the user feedback function,users are continuously received,which solves the problems of insufficient data in sentiment analysis and continuous optimization of the later model.In addition,the basic sentiment analysis function has expanded the comment rating prediction function,which can better analyze the impact of comment sentiment on the rating.Simultaneously,it also has new requirements for system authority management,model management,data visualization,and personal center.Corresponding solutions are provided.This topic systematically implements sentiment analysis and closes the system function,which has certain practical significance for expanding the comment sentiment analysis function in the current merchant system.