基于长短时记忆网络的COVID-19预测研究

Prediction of COVID-19 Based on Long term and Short term Memory Network

目的:探讨长短时记忆网络(Long Short Term Memory,LSTM)在新冠肺炎(Corona Virus Disease 2019,COVID-19)预测中的应用,实现对新冠肺炎更准确的预测。方法:以美国亚利桑那州新冠肺炎累计确诊病例数据为实验样本,采用基于长短时记忆网络构建的新冠肺炎预测模型进行预测,并分析不同因素对新冠肺炎预测模型预测效果的影响。结果:基于长短时记忆网络构建的新冠肺炎预测模型的均方根误差为3955.77,平均绝对误差为2959.85,平均绝对百分比误差为0.62%,均优于循环神经网络(Recurrent Neural Network,RNN)模型和反向传播(Back Propagation,BP)模型的预测效果。结论:基于长短时记忆网络构建的新冠肺炎预测模型准确率较高,对新冠肺炎疫情预测具有一定的实用价值。

Objective:To explore the application of Long Short Term Memory(LSTM)in the prediction of COVID-192019(COVID-19),so as to achieve more accurate prediction of COVID-19.Methods:With the cumulative confirmed case data of COVID-19 in Arizona as the experimental sample,the prediction model of COVID-19 based on long-term and short-term memory network was used for prediction,and the influence of different factors on the prediction effect of COVID-19 prediction model was analyzed.Results:The root mean square error of the prediction model of COVID-19 based on long-term and short-term memory network was 3955.77,the average absolute error was 2959.85,and the average absolute percentage error was 0.62%,which were superior to the prediction effect of recurrent neural network(RNN)model and back propagation(BP)model.Conclusion:The prediction model of COVID-19 based on long-term and short-term memory network has a high accuracy,which has certain practical value for the prediction of COVID-19.