Research On Sentiment Analysis Of Homestay Reviews Based On Text Mining

In recent years,as a new type of accommodation,homestay have gradually become a new choice for people to settle down during travel,and are favored by the public.At the same time,with the rapid development of Internet technology,people are more and more willing to comment on homestay services through online platforms to express their views,emotions and attitudes.These comments include influencing the decision-making of homestay managers and people's choice of homestay consumption of important information,which has great mining value.How to effectively identify,refine and analyze users' emotions and needs from the homestay reviews,and then provide information services for enterprises or consumers,is the starting point of this paper.This paper mainly applies the text mining technology to the emotional analysis of Guizhou homestay reviews,and studies the comment data on the platforms such as Ctrip.com,Xiaozhu.com and Qunar.com.In order to solve the problem of low recognition rate of users' emotion in new words,a fusion method of SO-PMI and Word2 Vec is proposed.This method takes full account of the co-occurrence and semantic association between words.On the basis of the original emotion dictionary,Guizhou homestay domain exclusive dictionary is constructed.In the process of experiment,the word frequency-inverse document word frequency technology is used to select emotional seed words,and the gradient descent method is used to assign different weights to degree adverbs to improve the accuracy of emotion recognition.It is found that the affective analysis performance of domain dictionary is improved to some extent,which is better than the dictionary expansion method based on SO-PMI and Word2 Vec respectively.In negative comments,F characteristic value increases by5.4% and 6.5% respectively.At the same time,this paper uses domain dictionary to mine the emotion of homestay reviews,and analyzes the relationship between users' emotion distribution and user evaluation level in different regions of Guizhou,and presents them by visualization methods.In addition,with the help of LDA model to complete the topic mining of emotional classification,the advantages and disadvantages of Guizhou homestay,that is,positive and negative themes,are analyzed in detail.Using the formula to calculate the user satisfaction of Guizhou homestay,and understand the differences among different regions.Finally,the detailed information of Guizhou homestay is counted and coded,a proprietary data feature set is constructed,and word-of-mouth classification and prediction are performed based on machine learning methods such as random forest,naive Bayes,and support vector machines.The experimental results show that the overall effect is good,which is an effective method to recommend high-quality homestay to consumers.