Short-term Forecast and Analysis of Mass Incidents Based on Time Series Model

摘要

In recent years, the high incidence of mass incidents has seriously threatened the stability and harmonious development of the society. Effective forecasting is very significant to the prevention of and response to mass incidents. Based on the Autoregressive Integrated Moving Average (ARIMA) model and Markov Switching-regime (MS) model, this paper forecasts the number of mass incidents in mainland China in short term. Through the residual error diagnosis of the fitted results and the error analysis of the forecasted results, it is found that the two-regime Markov switching model can achieve a more accurate prediction. The results show that the trend of mass incidents can be divided into two regimes: the stable stage and the peak stage, and the switching between the two stages can be explained by the third-party threats like the economic crisis, political events, and changes in the international situation.