As one of the representatives in social media,Weibo has become an important channel for people to express and disseminate opinions and emotions with its characteristics of convenience,communication,low threshold and high interactivity.By quantifying the dynamic influence of Weibo public opinion,we can analyze the evolution of public opinion at different time points,better grasp the attention of Weibo users to the current public opinion events,and discover the evolution law and potential risks of public opinion,thus for the government,Enterprises and other entities control,guide public opinion and develop public opinion early warning mechanisms to provide decision support.Based on the interaction behavior-microblog content-user emotion level,this paper proposes a public opinion influence measurement algorithm based on interaction behavior and emotion difference.Firstly,based on the user’s interaction behavior,the principle of TF*PDF algorithm is used to quantify the influence of public opinion interaction caused by different interaction behaviors.At the same time,the influence of public opinion trend on public opinion influence is considered,and the interaction behavior and public opinion trend are proposed.The lyric interaction influence algorithm;secondly,from the user’s emotional level to define the lyric emotional influence,using Bayesian plain classifier and other methods for sentiment analysis,calculate the microblog user’s emotional propensity value and use its variance to describe the current lyric sentiment difference The degree,referring to the calculation method of information quantity,quantifies the amount of information about the potential emotional influence caused by the conflict of emotional differences,and proposes the algorithm of affective influence based on emotional differences;finally,the comprehensive influence and emotional influence of public opinion The behavioral and emotional aspects are combined to quantify the apparently influential influence of public opinion.Taking the microblogging lyric event "DiDi drop down the windmill" as an example,the empirical analysis is carried out to verify the algorithm proposed in this paper.At the same time,the lyrical dynamic evolution analysis strategy of comprehensive interaction behavior and sentiment orientation is established,and the evolution characteristics of public opinion on the interaction behavior and emotion level in different evolution stages are analyzed.The empirical analysis proves that the dynamic influence measurement model proposed in this paper can identify the dominant influence based on interaction behavior and the invisible influence based on user emotion,and provide an effective tool for the dynamic evolution analysis and influence power of public opinion. |