Evolutionary Prediction of Nonstationary Event Popularity Dynamics of Weibo Social Network Using Time-Series Characteristics

<div>Popularity evolution prediction of social event: “Chengdu female driver was beaten by changing routes.” (a) Prediction accuracy comparison. (b) Cumulative relative error analysis. The predicted result of the NSEPP model with explanatory variables is the closest to the actual evolution. Social intensity can significantly reduce the cumulative error.</div>

Discrete Dynamics in Nature and Society

fig9

Figure 9

Figure 9: Evolutionary Prediction of Nonstationary Event Popularity Dynamics of Weibo Social Network Using Time-Series Characteristics 