Comprehensive Evaluation Of Network Reputation Based On Sentiment Analysis In Grade 5A Scenic Spots In Jiangxi

The rapid development of mobile Internet and online payment,so that the purchase of tourism this experiential consumption,gradually transferred from the line to the online.This also makes the major tourist website have become an important platform for visitors to share the travel experience and published evaluation of scenic spots.Thus the formation of the network reputation,has become one of the important factors affecting the potential tourists travel intentions.Similarly,the network reputation is also an important source of information for scenic area managers to grasp the image of the scenic spot and timely adjust the management plan of the scenic spot.Therefore,it is of great significance for the scenic area managers and the potential tourists to explore the scenic spot.In this paper,six 5A scenic spots in Jiangxi province were analyzed,and a comprehensive evaluation system was established by using the method of feature sentiment analysis.In this paper,use of the web crawler tools,from Ctrip.com,LY.com and Lvmama.com three major tourist website to obtain the 6 5A scenic spots of Jiangxi province from 2014 to 2016 in the year of the 23300 comments.With the R of the original data cleaning and segmentation.Do dependency parsing of the segmented data using the LTP-Cloud Dependency Parser component,From the results of the analysis,we find out the 30198 pairs of words which are composed of the characteristics of scenic spots and the emotional attitude of tourists,There are 3340 feature words and 1444 emotion words.In order to establish the comprehensive evaluation index system of scenic spots,we use R to build a LDA topic model and extract eighteen common topic in the comments.Combined the topics by related words,establishing the index system of comprehensive evaluation of scenic spot network reputation.Then according to the words of the eleven third level index,making the results of dependency parsing are classified into eleven third level index,with the sentiment analysis and the word similarity calculation,we can judge the emotional tendency of the features and sentiment words in pairs,and get the favorable rate of the eleven third level index of the 6 scenic spots.In order to carry out the comprehensive evaluation of the scenic spots network reputation,we construct the pairwise comparison judgment matrix based on the features and sentiment words in pairs of eleven third level index,by using the AHP model,the relative weights are calculated from the bottom layer.the six scenic spots were evaluated by the praise rate of eleven third level index and we can obtained the comprehensive praise rate of six 5A scenic spots.From the first to the sixth were: Mount LongHu scenic area,Mount Sanqing scenic area,Mount Jinggang scenic area,Expo Area of Acient Kiln & Folk Cultures in Jingde,Jiangwan Scenic Area of Wuyuan,Mount Lu scenic area,and the 6 scenic spots in the tourism resources and tourism overall performance is very good,but in the consumption and basic conditions is poor.Finally,according to the conclusions,12 suggestions are put forward to promote the construction of the network reputation of 5A grade scenic spots in Jiangxi Province.