Design And Implementation Of Microblog Opinion Leaders Monitoring System

With the prevalence of microblog social networking platform in China, more andmore people like to express and exchange their experiences or point of view atmicroblog platform. Microblog has become an important origin and fermentation ofnetwork public opinion. More and more attentions are being paid to the microblogpublic opinion in China and at abroad, a lot of manpower and resources are investedon the microblog media monitoring. Opinion leaders play an important role in theformation and development process of the microblog public opinion, so it isnecessary to pay attention to the influence of opinion leaders, which will help toexpand monitoring work of microblog public opinion.Microblog platforms generate enormous information every day which is shorttext and has the characteristics of fragmentation so that it brings difficulties to themicroblog public opinion monitoring. This thesis focuses on public opinion leadersand designs a crawler which can grab information of microblog opinion leaders. Thesentiment analysis of text information can effectively identify the views andstandpoints, which can be the reasonable guidance for public opinion and establish ahealthy and orderly environment of public opinion. The main work of this thesis is asfollows.Firstly, the thesis explores the formation process of public opinion and analyzesthe characteristics of the microblog opinion leaders, then studies the domestic andforeign research condition and the framework of the monitoring system related to thistopic，and points out the deficiencies in the field of research and proposes the solutionof this thesis.Next is to highlight the work of microblog crawler and text orientation analysis.Through using the API and some crawling strategy, the microblog crawler is designedand implemented successful, which provide data support to the work of short textanalysis. According to the correlation field characteristics of opinion leaders, thisthesis constructed a ontology combining HowNet and areas dictionary, base on whichgives a microblog short text tendency analysis method, analyzes the tendency ofopinion leaders effectively. Finally, a prototype system is designed and implemented base on the work above,which can show the public opinion of microblog opinion leaders to the user. Charttools of system demonstrate the tendency trend of opinion leaders clearly, and itworks with the data to support the monitoring of public opinion.