Research on Emotional Tendency of Catering Service Evaluation Based on LDA Topic model

Abstract

To enhance consumers' dining service experience, this article constructs a mathematical model for emotional tendency analysis based on the text content of dining service evaluation. Firstly, clean the data of the comment text, and then use the jieba library to segment words. After removing the stopping words, the positive and negative parts are obtained. Secondly, use the counter library to calculate word frequency. calculate the number of emotional words, degree words, exclamation marks, and negative words, and classify and summarize them. Finally, sum up to obtain the total score of the comment, and use the LDA theme model based on various emotional feedback to obtain the overall strengths and weaknesses of the catering merchant. Research has shown that the advantages of merchants who gain the most positive emotions are: good taste, affordable prices, etc. The disadvantages of catering merchants with the most negative emotions are: slow delivery speed, incomplete preparation of tableware, etc. The improvement strategy is to accelerate delivery speed, patiently serve them, and make consumers eat well.