Design Of Tourism Comment Text Information Mining And Visualization System

In recent years,with the introduction of the "Internet + tourism" model,tourism research has gradually become a hot spot.Traveling tourists obtain relevant information about tourism destinations through the Internet.Tourist attraction comment text is the primary information that tourists pay attention to and assists tourists in their travel decisions.Through the mining of tourist attraction comment text,tourists can further understand the traveling attractions.The mining of tourist attraction comment text is the best help to reduce the difficulty of travel.It is difficult to find the large number of popular tourist attractions and comments in the text,which is difficult to express in real time(1)The text data set of tourist attraction comments is constructedThere are few scenic spot comment text data sets.The experimental data set of this paper crawls 110000 comment texts of some scenic spots in Beijing on Ctrip.After crawling,it carries out text preprocessing and manually marks the emotional tendency of comment texts,and finally constructs 110 k scenic spot comment texts and 50 k emotional comment texts,so as to lay the foundation for later experiments.(2)A topic extraction model integrating semantic information is constructedAiming at the problem that the traditional topic model ignores the contextual semantic association of scenic spot comment text,word2 vec is used to learn the contextual semantic information of scenic spot comment text,and the trained semantic information is integrated into LDA topic extraction model.By learning semantic information,the lexical dependency between texts is improved,and pyldavis visualization technology is adopted,The results show that the topic analysis model integrated with semantic information has a good effect on the comment text of tourist attractions.(3)An affective analysis model integrating attention mechanism is constructedFor the comment text of tourist attractions,in order to better extract text features,this paper uses Bert + bi LSTM + attention model for text emotion analysis.Firstly,the word vector is obtained through the Bert model.Secondly,the advantage of extracting the global features of tourist attraction comment text by using the bi-directional longterm and short-term memory network(BI LSTM)integrating the attention mechanism is used to improve the effect of emotion classification.The experimental results show that the model proposed in this paper has better effect and achieves the best results in various evaluation indexes.(4)A prototype system of text visualization based on scenic spot comments is builtThe system presents the theme analysis,emotional tendency and keywords of scenic spots in a visual way according to the travel scenic spots,so as to provide convenience for tourists to choose travel scenic spots.