Research On Internet Memetic Evolution

The availability of data in online social media and human online behavior creates unprecedented opportunities for exploring human network cultural activities and social phenomena on a larger scale.In this case,one of the most challenging issues is studying the evolutionary dynamics of a series of memes such as thought,information,knowledge,and rumors.And in the process of diffusion and evolution,only a few ideas and information are widely disseminated.Therefore,these important issues are which characteristics make these “minorities” succeed,and how these successful memes evolve.In view of the above issues,the main work of this paper is as follows.Firstly,this paper proposes an agent-based internet meme diffusion evolution model for online users' behavior data;it is named as OUBD model(Online User Behavior Data Model),and explores the internet meme evolution mechanism.Adopting user entropy and overall entropy to quantify the relationship between user attention and overall information diversity through information entropy theory during data collection time;using maximum information path similarity measure,conditional probability,and Pearson correlation coefficient to analyze the interaction between user memory(interest)and online behaviors of users on the Internet;the statistical characteristics of experimental data,such as the meme life cycle,the popularity of memes,and the probability distribution of user activities are explored;on the basis of the above three issues,the Internet meme diffusion evolution model is proposed,which includes the memory mechanism and the way agents work,and analyzes the influence of the degree of Internet meme competition in the model on the statistical characteristics of Internet memes.Secondly,this paper proposes similarity of Internet memes,which is one of the internal factors for the development and evolution of Internet meme popularity.The similarity between Internet meme labels and Internet meme content is calculated by Latent Semantic Indexing and Jaro-Winkler distance.To achieve a single quantitative estimation among Internet memes through linear combination of similarity between Internet meme labels and Internet meme content;the relationship between Internet memes similarity and popularity evolution was analyzed based on Spearman correlation coefficients and other methods.Thirdly,the study found that no matter how the internet meme diversity changes,the user's attention span remains unchanged;the user's memory(interest)significantly affects the user's later behavior;the statistical characteristics of the experimental data are heterogeneous,and show power-law characteristics.One of the reasons for the heterogeneity of statistical characteristics is the degree of competition in the OUBD model;the similarity of some Internet memes correlates with its popularity evolution to a great extent.