Design And Implementation Of Commodity Reputation System Based On Sentiment Analysis Of Consumer Reviews

With the development of the Internet,massive amounts of transaction data will be generated in the process of e-commerce transactions,among which consumer product review texts contain great commercial value.At present,the number of product review texts is increasing geometrically,causing consumers to generate information overload problems.At the same time,the product review text has the characteristics of short text,many interference information,irregular writing,and complex expression,which also increase the difficulty of sentiment analysis.Mainstream e-commerce platforms use simple scoring mechanisms to make use of review information to a certain extent,but they still lack fine-grained mining of review information.Therefore,the establishment of a product word-of-mouth system based on consumer review sentiment analysis can fully mine review information and solve the problem of information overload.This article strictly follows the software development process for demand analysis,system design,system implementation and system testing.Use J2 EE technology,MVC development model,SSM framework,etc.to realize various functional modules,including data capture module,data storage module,sentiment polarity analysis module,fine-grained word-of-mouth mining module and information display module.At the same time,this article applies the text sentiment analysis method to the analysis of word-of-mouth data,so as to obtain consumers' coarse-grained and fine-grained evaluations of commodities.The core functions of this system can be divided into two aspects.On the one hand,the BERT-RCNN model is used for emotional polarity analysis to obtain coarse-grained word-of-mouth products;on the other hand,the LDA topic model algorithm is used to analyze product reviews from terms of word level and characteristics.Word-of-mouth mining is performed at word level and opinion word level to obtain fine-grained word-of-mouth products.At the same time,a visualization system was built to allow users to more intuitively and quickly understand the reputation of the product,successfully solved the problem of information overload,and enabled consumers to obtain a better shopping experience and higher shopping efficiency.The product word-of-mouth system based on consumer review sentiment analysis designed and implemented in this paper has been developed and tested.The average acquisition time for every 1000 comments and opinions of the CNAS-certified third-party testing agency is 159.81 seconds,compared with current e-commerce.The platform's word-of-mouth display mechanism has been significantly improved,which has improved the shopping efficiency of consumers and optimized the shopping experience.