Design And Implementation Of An Emotional Analysis System For Commodity Reviews

With the rapid development of the network in recent years,the social system has become an interconnected society,and the way of communication,the way of travel,the way of learning,and the way of life are all innovating.Now,the geometric growth of data makes it unrealistic to collect the user's attitude(emotion)simply through manual screening.Now,it is urgent that an emotional analysis system,such as an emotional analysis system,can be quickly and better to extract the commodity reviews,and analyze the emotions in these comments and make a good job of their products.The standard of promotion makes the consumer willing to pay the bill.The purpose of this study is to establish an emotional analysis system for commodity reviews.This system is used to analyze the comments of the users and to get the advantages and disadvantages of the goods from the emotional words.It can be divided into three parts to construct emotional analysis system.In the first part,we crawl and deal with the commodity reviews.In this part,first,crawling technology is used to crawl the review data of computer goods on the Internet,and then the data are tagged with "0" and "1",and the positive and negative samples are taken out after marking.In the second part,feature selection and feature dimension reduction methods are studied.Feature selection includes word features,two collocation features and stuttering participles.The feature of the selected features is reduced by statistical analysis algorithm,only the feature of rich information is retained.Simplification is beneficial to improving the speed of the algorithm and increasing the efficiency of the algorithm.In this part,the attribute feature dimension,attribute emotion and attribute weight concept are put forward,and the concept of attribute feature dimension,attribute emotion and attribute weight is put forward in this part.In order to improve the accuracy of classification algorithm,the accuracy of classification algorithm is improved.In the third part,we use the machine learning algorithm to train these feature information.This paper mainly uses the naive Bayes algorithm,support vector machine algorithm and Knearest neighbor algorithm.After comparison,the results are analyzed.Finally,the support vector machine algorithm is selected as the final comment statement emotion analysis system algorithm.After constructing the sentiment analysis model,we need to verify its robustness.Using the machine learning algorithm as the main technology of the emotional analysis system,a large number of data are needed to optimize the training model.By increasing the amount of data,the classification effect will be better and better,and the accuracy of the classifier can be improved gradually.In this paper,the system is an emotional analysis of the review of computer goods.With the optimization of the algorithm and the training of high quality data,the classifier is getting better and better.Finally,the method can be applied to all aspects of commodity reviews and finally realize the purpose of emotional analysis for any commodity comment.