Sentiment Analysis For Tourism Review Based On Hadoop

With the improvement of people's living standards and the gradual emphasis on quality of life,more and more people tend to go out for tourism.With the popularization of smart terminals and the rapid developme nt of mobile internet,a large number of online travel websites have emerged,making people feel free to With the help of the Internet to obtain tourist information,book travel products,publish travel reviews,and share travel experiences.These review d ata are lack of effective processing and application on various travel websites.Traditionally,data processing and algorithms have not been able to meet the performance bottleneck caused by more and more data explosion.The analysis and processing based on big data cloud platform has become a success Trends.The text focuses on the current status quo and constructs a travel network commentary data processing platform that can analyze and process the review data in a timely manner.Due to the exploding growt h of online commentary data,Hadoop,a big data processing platform,was used for data storage and analysis.The algorithm design based on the Map Reduce parallel framework can effectively deal with the data that cannot be solved by traditional data processi ng platforms.In the sentiment analysis,the review data is firstly preprocessed,and the emotional words and evaluation objects of the review data are extracted using dependency syntax analysis,and an emotional dictionary of travel evaluation data is constructed according to the travel review data to select features.The parallelization of NB,SVM,and KNN was modified,and propensity classification experiments were performed on the review data to compare the accuracy,recall rate,and F value.In determining the sentimental orientation,theme words under various themes are realized through the LDA subject model.In this paper,the natural language processing is used to analyze the dependency data of the evaluation data,so that the review data is denoised prior to the LDA subject model,which makes the emotional vocabulary and evaluation objects appear in the topic words.Finally,two semantic constraints are proposed.The topic model is clustered according to the rules.The results show that the method is effective in the sentiment analysis of comment data.