Research And Implementation Of Hot Spot Detection And Tracking System In Network Public Opinion

With the continuous development of network technology,more and more people express their opinions and get information through the network media,so more and more people pay attention to the Internet public opinion.The Internet public opinion is spread through the Internet,so it inherits the characteristics of concealment,openness and infiltration of Internet,at the same time,it contains rich information and it has a very large amount of information.Due to these reasons,collecting network public opinion information by humanity is very difficult,and the network public opinion monitoring system was created.The Internet public opinion system gather information in real time,then process and analysis these information,at last the final results are displayed to the user in a form which can easily understand.So the internet public opinion is useful to help users dealing with crisis of public opinion correctly.This paper summarizes the existing network public opinion hot spot detection and tracking method,did some research of the key technology like the content of topic detection and tracking technology at first.By understanding these technologies,grasp the key knowledge of the development of hot spot detection and tracking system.Then,it studied the parsing technology of web content.According to the different characteristics of the news websites,we extract a lot of information which can be used in next step,and we save these information for feature using.At last,it explains the discovery technology of topics,the evaluation technology of heat of topics and the tracking technology of topics in detail.And it uses the vector space model to represent the information text,including the feature items set,an improvement of the weight calculation formula and a determination of the dimension of features,and it uses the improved similarity algorithm single-pass to cluster,so the clustering results will be more accurate.In the research of heat of topics,the paper calculates the heat from the media attention and user attention through counting the number of reports and comments of the topic,then combines the results to get the final heat of the topic.At last,this paper does some research in the tracking technology,through the analysis of the variation tendency of the number of reports and comments,it shows the change of the topic's degree of attention,and through the analysis of the variation tendency of the number of the sub topics' reports of every topic,it shows the migration scenario of the topic center.In this paper,a hot topic of network public opinion detection and tracking application system is finally realized preliminarily.This system has realized the function of web page pretreatment,hot topic discovery and analysis,hot topic tracking,and it has some certain practical significance.