Study On The Influence Of Web Reviews On 5A-Class History And Culture Spots' Incomes Based On Text Mining

Web reviews have been proved to be one of the important factors affecting consumers' consumption decisions.In terms of the scenic area,web reviews objectively reflect the true feeling of the tourists on tourism and service,which influences the potential tourists travel decision-making and scenic operations,and scenic incomes can illustrate operation state basically,thus this paper mainly discusses web reviews how to affect scenic incomes.Based on the text mining technology,taking5A-class History and Culture Spots as the research object and climbing web reviews from Lvmama.com,Meituan.com and Ctrip.com as the research material,this paper explores the impact of web reviews from the quantities and text contents on the scenic spots' incomes with the use of python3.7 software.On the one hand,this paper studies the factors affecting 5A-class scenic incomes,including web reviews' quantities,ratings,tickets,tourist arrivals,Baidu Index.After building multivariate linear regression model,the result proves that the relationship between web reviews' quantities and the scenic areas incomes is positive,and different types of scenic web reviews on incomes effect is different with maximum effect of history and culture scenic spot.In order to make the research more specific and targeted,this paper takes the 5A-class History and Culture Spots as an object.On the other hand,this paper studies the influence of web reviews content on the 5A-class History and Culture Spots' incomes,providing a reference for other types of scenic spots.Specific as follows: the first step uses the LDA model to extract the topic information in the web reviews,and explores the tourists likes and dislikes of5A-class History and Culture Spots and focuses on characteristics of potential information.Combined with tourist perceptions,it builds a 5A-class comment text classification standards,5 first classificationsconsisting of experience,scenic consumption,scenic resources,scenic services and scenic management,12 secondary classifications.Second step builds a semi-supervised support vector machine classifier according to the first classifications,then categorizes each comment,and values emotional tendencies grades based on the emotional tendency of dictionary.It founds that there are more comments about tourists' experience,scenic resources,scenic consumption,and the tendencies of emotional expression are positive.Final step studies the impact of different classifications of web reviews on scenic incomes,and finds that the experience,scenic resources,and scenic services have a significant impact on scenic incomes.According to the analysis conclusion,this paper puts forward to four suggestions--use scenic resources effectively and deep the connotation of experience;Improve the quality of service staff and optimize the experience;Improve the management level of scenic spots and upgrade the management concept;Guide scenic consumption orderly and optimize scenic management reasonably--for the sake of promoting the scenic operation and improving scenic incomes.