Design And Implementation Of Commodity Evaluation System Based On Emotional Analysis

While online shopping is convenient for people's life,there are also some problems.Because online shopping can't touch the goods as offline shopping,and all the information of the goods is given by the merchants,this causes the information inequality,which easily leads to the users to buy fake goods,defective goods,or goods that don't meet their expectations,causing certain losses.Although users can get valuable information by viewing product reviews,with the continuous accumulation of product reviews,the efficiency of this way of information acquisition is getting lower and lower.Aiming at the problem of excessive accumulation of commodity reviews,the key point is to extract valuable information from the mass of commodity reviews.This paper designs and implements the commodity evaluation system based on emotion dictionary and SVM through the research of existing emotion analysis methods,analyzes and processes commodity reviews by using emotion analysis methods,and provides users with objective and comprehensive shopping reference information.The main work of this paper is as follows:(1)Data collection.Based on the analysis of e-commerce platform page,this paper designs a crawler based on scrapy framework,which can crawl the information of corresponding products from e-commerce platform.In this paper,hundreds of thousands of data are collected from e-commerce platform by using the crawler.(2)The research of emotion analysis method.In order to extract valuable information from commodity reviews,this paper studies the existing affective analysis methods,and adopts two affective analysis methods based on affective dictionary and SVM to deal with commodity reviews.In the research of emotion dictionary method,based on the general emotion dictionary,this paper uses word2 vec model to expand by calculating word similarity,constructs a special emotion dictionary,and designs the corresponding scoring rules.In the research of SVM method,this paper first uses the combination of dictionary and user score to build the training set,then uses word2 vec model to generate the word vector of feature words,and in order to better show the importance of words in the word vector,this paper uses the way of word vector weighting to improve the generated word vector.(3)The design and implementation of commodity evaluation system.In this paper,using the framework of flask + Vue,adopting the development mode of front-end and back-end separation,the design and implementation of commodity evaluation system is completed with the emotional analysis method as the core.Users can get the results of emotional analysis,product characteristics analysis and other information through interaction with UI interface,so as to know the overall evaluation of the product and the advantages and disadvantages of the product characteristics.Provide users with objective and comprehensive shopping reference information.