Research On Some Key Techniques Of Burst Topic Public Opinion Analysis In Microblog

With the rapid development of social media platforms,social media have become the main platforms for obtaining and spreading information,which gradually replace the traditional media.However,while people use microblog and other social media to facilitate the dissemination of information,social media have become important platforms for the emergence and dissemination of burst topics.Different from the traditional media,burst topics in microblog cannot be restricted by time,space,which makes burst topics detection and mining analysis more difficult to be controlled.When burst topics involving sensitive information in microblog occur,it will become a wide influencing social event and endanger the security of the whole society if negative public opinion cannot be controlled.As a result,opinion mining analysis for burst topics in microblog is paid more and more attention to by academia and industry.In this paper,we take the most representative microblog platform as the researching object and aim to research the methods and techniques of burst topics detection and mining analysis.Because microblog has a large amount of data,serious information fragmentation,spotty user quality and fast information transmission characteristics,artificial real-time monitoring cannot detect and mine burst topics in microblog effectively.Therefore,for the source of social media public opinion,such as microblog,how to automatically detect burst topics oriented realtime microblog stream,mine and analyze detected burst topics,effectively to prevent the microblog public opinion crisis and correctly guide the microblog public opinion are the key problems of the social network public opinion security.Our paper mainly focuses on the perspective of burst topics detection and mining analysis in microblog,mainly conduct the research from the following aspects:First,in order to detect burst topics real-time,we propose microblog pre-processing methods for burst topics detection in microblog.For the method of keywords pre-processing,we propose a novel approach to detect burst keywords based on social trust and dynamics model.We adapt basic notions of dynamics from physics and model keywords bursts as momentum change of the keywords.On the analogy of physical dynamics model,this approach defines mass as the trustworthiness of user and position as the frequency of keywords.We compute each keyword's burst value by using Moving averageconvergence/divergence(MACD)and determine whether it is a burst keyword in a given time window.The experimental results on large-scale Sina microblog dataset show that the proposed approach can avoid detecting fake burst keywords.For the method of microblog user pre-processing,in order to detect bots in follower markets in microblog and reduce the impact on burst topics detection,we propose an effective approach for bots detection based on interaction graph model.Based on empirical analysis of collected dataset,we build an interaction graph model based on user interaction and design robust interaction-based features.We conduct a comprehensive set of experiments to evaluate the proposed features using different machine learning classifiers.The results of our evaluation experiments show that our proposed features are more effective to detect bots compared to other existing state-of-the-art approaches.The methods of microblog burst topics detection and bots detection laid the groundwork for follow-up studies of burst topics detection and mining analysis.Secondly,due to burst topics related to negative public opinion are usually social events,we propose an online burst events detection framework to detect burst topics oriented realtime microblog stream.In this framework,an efficient microblog message store and update model based on the sliding time window is conducted.Burst messages detection algorithm that can adjust the threshold adaptively is used to detect burst messages.Burst messages which are post by bots are removed.Combined with burst keywords and event features,an online incremental clustering algorithm is used to cluster burst messages and detect burst events.Experimental results in the realtime microblog message stream environment show that our framework can be used in online burst events detection and has higher accuracy compared with other approaches.Thirdly,for key users mining in burst topics,we consider the effect of key users in burst topics for public opinion diffusion and propose a framework that can detect community pacemakers in burst topics.In the framework,a burst topic user graph model is proposed,which can represent the topology structure of burst topic propagation across a large number of Twitter users.Based on the model,a user community detection algorithm based on random walk is applied to discover user community.For large-scale user community,we propose a ranking method to detect community pacemakers in each large-scale user community.To test our framework we conduct the framework over Twitter burst topics detection system.Experimental results show that our method is more effective to detect the users that influenceother users and promote early diffusion in the early stages of burst topic.Finally,for burst pattern mining in burst topics,we investigate the problem of mining burst patterns of burst topic in Twitter.A burst topic user graph model is proposed,which can represent the topology structure of burst topic propagation across a large number of Twitter users.Based on the model,hierarchical clustering is applied to cluster burst topics and reveal burst patterns from the macro perspective.Frequent sub-graph mining is used to discover the information flow patterns of burst topic from the micro perspective.Experimental results show that several interesting burst patterns are discovered,which can reveal different burst topic clusters and frequent information flows of burst topic.