Study On Micro-blog Public Opinion Modeling And Developing Trend Prediction

The development of information technology and internet have changed people's life greatly.The Internet has gone deep into all respects of the society and has greatly influenced people's work,life,study and amusement.The emergence of it has had a tremendous impact on traditional information media.As an extension of public opinion in society,the public opinion on micro-blog platform plays a pivotal role.Due to the release of information on micro-blog is timely,arbitrary,and the dissemination of information is very rapid,making the information on micro-blog is both real and false.Some people use micro-blog as a tool for spreading rumors,which has caused a great negative impact to social ecological environment and people's normal life.Therefore,it is of great practical significance to study the trend of public opinion on microblog platform.Before we use algorithm to predict the trend of hot topics on microblog,we need to collect relevant public opinion data.In this dissertation,we get data from Sina microblog,and find the hot topics from the obtained data.The steps are as follows:(1)The method of combining web crawler with micro-blog API is used in data acquisition;(2)After the microblog text is preprocessed,the clustering algorithm is used to get hot topic;(3)According to the characteristics of micro-blog,we can extract the characteristics of microblog's public opinion trend,and form the public opinion prediction experimental data.At present,RBF neural network is one of the most popular and successful among neural network,it has the advantages of simple structure,strong plasticity,and has the global optimum approximation ability and good generalization ability.But when it comes to public opinion,the accuracy of the prediction results of RBF neural network is closely related to the center of the radial basis function,the variance(width)and the weight of the hidden layer to the output layer.Therefore,in traditional RBF neural network,the choice of the parameter limits its application in prediction of network public opinion.By comparing the experimental results of three models,it is found that the model which is optimized by improved gravitational algorithm has the best precision.Therefore,the new algorithm model can better predict the development trend of the microblog topic.The prediction results are conducive to government's monitor and guidance of public opinion,and also conducive to social harmony and stability.