Research On Tourism Information Mining And Visualization Based On We Media

"We media" is a new concept emerging with the development of web technology,mobile Internet technology and intelligent terminals in the past decade or so.It is a kind of independent new communication method for the common people,as communicators,to transfer their knowledge,opinions and events to other audiences."We media" is characterized by common people,low threshold,fast transmission and strong interaction,while online "we media" tourism information has great mining value and application value.Tours in the backdrop of the mainstream tourism,most visitors will choose free way to travel,most visitors will be after the end of the travel behavior on the travel portal interaction module to share their travel experiences,leads to "we media" in the form of a variety of online travel information is exploding.Based on the media's online travel portal website contains many kinds of tourism information,including the resort and its attribute information,the various comments on data,the number of online travel,the travel quiz,and each type of tourism data contains a variety of various types of attribute information,for example,time information,text information,etc.However,the surge of online travel information is not a good situation for self-service tourists,tourism practitioners,etc.,which is information "overload" for them.how to dig out useful tourism information points from massive self-media online travel information,and to mine this information from multiple dimensions,so that this part of information can be effectively utilized is an important value of tourism information mining.The development of computer information technology,graphics and image processing technology has also changed the traditional way of visualization and made great progress in modern visualization technology.As for the results of information mining,it is necessary to find an appropriate way for visualization,so as to help people enhance their cognition and give full play to the important role of visualization in data understanding.On the other hand,the research results of this paper can be effectively applied to the modern smart tourism system to save time and provide targeted Suggestions for tourists when choosing destinations.Based on "we media" tourism information,this paper takes Shanghai as the research area and makes an in-depth study on the acquisition,mining and visualization of tourism information from the perspective of theoretical research and practical application value.The research content of this paper mainly focuses on the following four aspects:(1)Different methods were designed to collect and preprocess the self-media travel information.Taking Shanghai as the research area,through the previous research,statistics and comparisons were made on the data types,numbers and online users of the travel guide messages in the major travel portals,and finally the data quality was higher and the users were more.The Ctrip website serves as the data source for this article.Based on Python web crawler technology,the collection methods and processes of different data types are designed,and the raw data is pre-processed to provide data support for subsequent tourism information mining and visualization.As of July 2018,a total of 5,293 Shanghai tourist attractions,Shanghai Disney Resorts,a total of 72,685 reviews,3,000 online travel notes,and 3018 travel quiz data on Shanghai were collected.After pre-processing,only 4,302 tourist attractions,63,115 reviews,2,100 travel notes,and 700 travel destinations were retained.After pre-processing,it is stored in the MYSQL database.(2)Mining the collected tourism information,and designing different mining methods according to different data types.First of all,the evaluation time was excavated.For the Shanghai Disney Travel Review data,because the information is relatively structured,it can be directly analyzed and analyzed.The interannual variation of Shanghai Disneyland from 2016 to June 2018 is analyzed.And the amount of monthly changes throughout 2017.For the review text itself,according to its data characteristics,the classification process and evaluation indicators of the review text were designed by means of machine learning.In the process of experiment,through text preprocessing and text word vector representation,it is automatically classified,and the experimental results are analyzed.The influence of different classifiers and different sample sizes on the classification results is obtained.For the short text of question and answer,this paper directly calls Bonson NLP's multi-text clustering to get top10 of relevant questions and answers in Shanghai,which provides data support for subsequent visualization.According to the online travel records of Shanghai tourism destinations,the excavation process and method of hot tourism place names were designed.In the experiment process,the hiragana word segmentation based on hidden Markov model was used to identify the place names,and the ATF*PDF model's place name vocabulary weights were used.The allocation method performs a weight assignment calculation on the hot place names.Finally,the accuracy of the extraction results is analyzed.Different geographical names are classified based on the extraction results and spatialized by Arc GIS.(3)Design and Develop a tourism information visualization platform based on we-media tourism information mining results.GIS secondary development technology,combined with HTML+CSS front-end development technology and Gaode Map API,Echarts and other visual components of the platform construction and development.Through this platform,massive and disordered tourism text information is displayed in a way that is easier to be fully understood,in the form of map and text.