Microblogging Information Analysis Based Design And Implementation Of Prediction For Tourist Attactions Popularity

Microblog as the most popular social networking media, it has a certain role andsignificance of the predicted heat of tourist attractions, not only can help users chooseto provide decision-making reference value when the fun attractions, can also helpbusinesses to provide users with personalized Recommended tourist attractions, andeven prevent the occurrence of unexpected events have a certain role. To achievetourist attractions heat prediction system, a number of issues need to be addressed asfollows: First, the amount of data microblog database is explosive growth; the largeamount of data has gone beyond the traditional techniques of data processingcapabilities. Second, the traditional keyword extraction algorithm only considering therelationship between the key words and the number of it appears’ texts, ignores thekey words in a category of distribution, leading to decreased accuracy problems ofmicro-blog commercial word extraction. Third, in order to obtain valid key users,need to exclude the interference of zombie powder and advertising users. Fourth, thecurrent studies are directed at the contents of the past or the current record ofinformation did not do projections, and therefore need to design the relevantalgorithms.Faces of the above problems, firstly, designed algorithm to eliminate interferencefrom advertisers and influence of zombie powder to get effectively key users.Secondly, the mass microblog information has been classified by the relevant tourismmicroblog; followed by the travel-related microblog for Chinese word segmentation,and used improved TF-IDF function to calculate term weight. After the weight ofterms to be sorted, we can obtain the high frequency attractions vocabulary in acertain time. Then by analyzing the behavior of influence between users, combinedwith a key set of users and high tourist class set of keywords to analyze trends in thespread of computing tourist information, which can predict the heat of touristattractions. Finally, the system is ported to the Hadoop Distributed framework.Experimental results show that the system is feasible and effective, and hadoopframework for processing huge amounts of data quickly.