Research On Public Opinion Hotspot Monitoring Based On Wuhan City Message Board

With the rapid development of science and technology,the Internet has become an indispensable part of people's life.The public can speak freely on the Internet and exercise their rights of supervision,information,expression and participation.As a result,the profound influence of online public opinion on the society is in-evitable.Because the hotspots and outbreaks are always difficult to predict,and once happens,it will be hard to control,which makes the government pay more e.ffort to effective.ly monitoring the network public opinion.Meanwhile,it is nec-essary to find hot spots of public opinion,understand the emotional tendency of public opinion,accurate warning outbreak,and make adjustment according to the situation.And the government should give some timely and appropriately guides,in order to reduce the negative impact.At present,there are several ways to monitor public opinion hotspots,such as keyword extraction,document topic extraction,text classification,emotion anal-ysis,and so on.The monitoring process requires multiple algorithms from natu-ral language processing and machine learning.Common methods include Chinese participle of jieba,Word2vec text information vectorization,JKD text similarity coefficient,k-means clustering,hierarchical clustering,support vector machine,etc.In order to listen to the voice of the people,the Wuhan municipal party commit-tee and government launched the" city message board" function,which was widely-used by the public.Among them,the number of messages left by the housing author-ity under the government module is the largest.Based on the monitoring methods of public opinion hotspots at home and abroad,this paper analyzes the difficulty in simplification from the perspectives of words,sentences and events,including pub-lic opinion hot spot excavation,analysis of public opinion and emotion,and public opinion event warning.In addition to conventional methods,in the monitoring of explosive events,the maximum cluster problem of graph theory is introduced to solve the problem,to ensure that messages of explosive events are correlated with each other,and to avoid the problem that the number of classes or the location of layers cannot be determined in the clustering algorithm.