Sentiment Analysis And Application Of Commodity Online Comments

The rise of e-commerce allows consumers to stay at home to be able to buy their favorite products,and it's a great convenience to customers.Web2.0 makes consumer publish reviews after buying goods on the network become a habit.Commodity reviews are the subjective feelings of consumers after using the product,to express the emotional tendencies of consumers,and compared to merchandise information provided by the business.Consumer reviews are more reliable.Use of sentiment analysis technology to mine emotional information in commodity reviews has become a hot topic in research today.Emotional information can help consumer make purchasing decisions.It can help businesses understand their competitive advantages and disadvantages,and can provide reference information for supervising departments to carry out supervision work.Emotional information contained in the product reviews has some influence on consumers' purchasing decisions.This impact will be reflected in the changes of the easy-to-get economic variables such as the sales situation of the goods and the price.Relationship between emotional information and sales of goods is the main content of this paper.This paper first introduces the basic theories involved in the process of text sentiment analysis,including the use of web crawler for data collection,word segmentation,part-of-speech tagging,text feature representation to preprocess online comments text and commonly used emotion analysis methods.On this basis,this paper puts forward the method of comment text emotion analysis.The word emotion analysis uses the method of combining the emotion dictionary with the point mutual information-information retrieval algorithm.When analyzing the commentary emotion,consider factors that influence the direction or degree of their sentiment tendencies.Do corresponding treatment of these factors in order to improve the accuracy of commentary emotion analysis.And the two criteria of consumer membership grade and commentary word number are taken as the basis for judging the quality of comments,and the influence of the quality of comments on the sentiment tendency is taken into account in the model of sentiment analysis.Then the results of sentiment analysis applied to the sales forecast,the AR model is used to forecast the sales volume of the goods.The sentiment factors and the sentiment factors in the quality of the comments are integrated into the autoregressive model,and.an emotional perception model for short-term prediction is proposed.In order to strengthen the persuasiveness,this paper uses the web crawler to collect the real book review information of three e-commerce online shopping platforms and use the corresponding evaluation index to validate the models improved.The result shows that the improved model performance with a certain increase.