Analysis Of Online Public Opinions Based On Micro-blog Data

With the rapid development of the Internet,micro-blog and some other social network platforms,because of their convenient way of circulation and rich content of information,become an important birthplace of the spread of social public opinion,people can express personal feelings and comment on social hot topics through micro-blog.Starting from the specific hot micro-blog topic,this thesis studies the features of the spread of the public opinion and predicts the heat changes of the spread of the public opinion.This article analyzes the emotion to the hot topic according to the user classification and the propagation of the heat changes.The main contents of this thesis include the following aspects:1)Data acquisition and preprocessing.After obtaining data from the official data website,the unstructured data needs to be preprocessed,including data cleansing,outlier analysis,repetitive data monitoring,data transforming and sampling,etc.2)Analyzing the law of the spread of the network public opinion based on the topic micro-blog.Studying the change trend of the public opinion over time by using the forwarding volume,the number of comments and the number of praises in the topic micro-blog data,and analyzing the outlier in the process of change for the reasons.Finally,effectively predicting the heat of discussion toward the public opinion through the establishment of the time series model.3)The type and emotion analysis of the users of the spread of public opinion.First,make a descriptive statistics analysis to the users of the spread of public opinion,including the users’ forwarding methods,the variance analysis between the member users and the certified users,etc;then through the users’ five variables to make effective clustering of the users’ types,and to analyze the users’ features(users’ behaviors)of different types;finally,through text mining and analyzing to the unstructured data,to extract and visualize the users’ emotional characteristics.Based on micro-blog data,studying the law of the spread of public opinion on the network is helpful for us to understand the development trend of the events,and then helps the government to monitor the public opinion and manage the vice information effectively.