为更加精准地拟合和预测流行病病毒的动态传播趋势,提出分数阶网络动力学分析方法.考虑病毒感染特性,应用流行病动力学方法构建易感—潜伏—感染—隔离—恢复—死亡—免疫(susceptible-exposed-infected-quarantine-recovered-death,SEIQRDP)模型,建立具有记忆效应的分数阶复杂社交网络SEIQRDP模型,应用最小二乘法和预测校正方法,对所提出的模型进行拟合和预测,以湖北、四川和安徽三个不同地区的真实新型冠状病毒感染数据为例进行验证.结果表明:与整数阶相比,分数阶复杂网络模型的均方根误差(root mean squared error,RMSE)值更小,能够更好地拟合真实数据.
Fractional-order Network Dynamics Analysis of Viral Transmission
In order to precisely capture and forecast the dynamic transmission patterns of epidemic viruses,a novel method for fractional-order network dynamics analysis has been introduced.Initially,taking the virus's infection characteristics into account,the suscepti-ble-exposed-infected-quarantine-recovered-death(SEIQRDP)model is formulated through conventional epidemic dynamics techniques.Subsequently,a fractional-order complex so-cial network SEIQRDP model that incorporates memory effects is developed,and the model is refined and predicted using the least squares method and prediction correction method.Finally,validation of the proposed model is conducted using actual infection data-sets from three distinct regions—Hubei,Sichuan,and Anhui.The findings reveal that the fractional-order complex network model outperforms integer-order models by exhibiting a reduced root mean squared error(RMSE)value and a superior ability to fit real data.
epidemic dynamicsfractional ordercomplex networkSEIQRDP model
万方数据
