Research On Public Opinion Emotional Situation Model Based On Information Measurement

With the rapid development and wide application of the network,the network has penetrated deeper into the human society.The public opinion based on communication also has the characteristics of networking.Due to the strong randomness,rapid spread,large range of influence,high potential harmfulness,and difficult regulation and control,reflecting the development trend and emotional tendency of Internet public opinion events validly and providing quantitative indicators for decision makers has become one of the most important needs in the current research field of Internet public opinion.This paper proposes a research model of emotional situation of public opinion based on information measurement,and introduces information entropy into the field of public opinion analysis,which is used as the basis for judging emotional situation.The research has practical significance.At present,the researches of scholars mostly use the quantitative characteristics of Internet public opinion text as the basis of calculation to judge the situation and sentiment of Internet public opinion.This paper introduces information entropy and its related concepts into Internet public opinion research and expands it,and proposes a research model of emotional situation of public opinion based on information measurement.The information measure in this model is consistent with Shannon's idea of information entropy.In particular,this model uses word frequency to approximate the frequency of word occurrence,and calculates the information entropy of a sentence using the information entropy calculation formula.For sentences with father information,this model calculates the mutual information relative to father information,which is subtracted from the information entropy of the sentence.Thus we get information entropy related to the topic but not repeated in the sentence,which is viewed as the contribution of the sentence to the total information entropy of the events.In this paper,the total information entropy of events is counted and used as the basis of judging the relative hotspots of events.The total information entropy of events is counted and used as the basis of judging the development trend of events.Now,the sentiment classification of Internet public opinion emotion research is still dominated by positive and negative categories.However,the dimension is too small to reflect the true emotional tendency and distribution of public opinion groups.This paper divides emotions into seven categories,which are ‘cheerful',‘fond',‘angry',‘sad',‘disgusted',‘fearful' and ‘shocked' in order to approximate the true emotional distribution of the public opinion subject. This paper expands the ontology database of Chinese emotional vocabulary of Dalian Polytechnic by using network dictionary and expression dictionary,and combines negative words,degree adverbs and sentence patterns to analyze the emotional situation of Internet public opinion through dictionary matching method.By calculating and counting the proportion and absolute value of each emotional component in the event time period,it can be used as the basis for judging the emotional distribution and emotional intensity of the event.In the research of refining the sentiment analysis of Internet public opinion,the research method based on the quantitative characteristics of the text is still used,and there are existing inevitably problems of lack of objectivity and authenticity.In order to solve this problem,this paper combines the Internet public opinion situation analysis model based on information entropy and proposes the Internet public opinion sentiment analysis method based on information entropy.This method uses extended hyperinformation in the model as the basis of text sentiment updating,and improves the objectivity and authenticity of sentiment refinement research.In order to verify the validity of the above models and methods,this paper uses crawlers to collect data of 15 events,which are from the July 27,2017 to August 16,2017,and the total number of Weibo data is 1.44 million.The model and method validation are carried out to verify the results and facts.The results are in good agreement and reflect the advantages over the quantitative characteristics.