一种基于GABP神经网络算法的新冠肺炎模型预测  

A Prediction Novel of Coronavirus Pneumonia Model Based on GABP Method

作　　者：刘志勇[1] 王小红[2] Liu Zhiyong;Wang Xiaohong(Xianyang Vocational and Technical College,Xianyang Shanxi 712000,China;Shaanxi Provincial Party School of the CPC,Xi'an 710061,China)

机构地区：[1]咸阳职业技术学院,陕西咸阳712000 [2]陕西省委党校,西安710061

出　　处：《科技通报》2022年第10期47-53,共7页Bulletin of Science and Technology

基　　金：咸阳职业技术学院2022年度教学改革研究项目(项目编号:2022JYA03)。

摘　　要：针对新冠肺炎疫情发展模型的预测问题,提出了一种基于改进遗传算法误差反向传播(genetic algorithms back propagation,GABP)神经网络算法。该算法通过对遗传算法中的交叉、变异、及适应度等函数进行了重新定义,在获得GABP神经网络最优参数的同时,提高了最优参数的泛化能力,进而缩小了GABP网络寻优的搜索范围,并进一步获得新冠肺炎疫情发展的精确预测模型。最后通过仿真,验证了该GABP神经网络对模型拟合预测的准确性。A genetic algorithms back propagation(GABP)neural network novel is proposed for predetermination the coronavirus pneumonia model.In the proposed method,the crossover,mutation and fitness functions are redefined,and then the optimal parameters of the GABP network are obtained.The search scope of the GABP network is reduced while improving the generalization ability of the strategy.Meanwhile an accurate prediction model for the development of COVID-19 can be obtained.Finally,the accuracy of the model fitting and prediction by the GABP neural network is verified with simulation.

关 键 词：遗传算法 神经网络 新冠肺炎 模型预测 参数更新 

分 类 号：TP291.9[自动化与计算机技术—检测技术与自动化装置]