基于LSTM模型对印度新冠肺炎疫情的预测

Prediction of COVID-19 epidemic in India based on LSTM model

自2019年新冠肺炎疫情暴发持续到现在,无论对国家、社会、还是个人都产生了巨大的影响。部分国家及时采取了一些应对措施,但是依然无法完全控制住疫情的传播。目前,机器学习算法被用来预测新冠肺炎疫情的发展。使用长短期记忆(long short-term memory,LSTM)模型对收集到的印度累计确诊病例数据进行分析并预测印度疫情的变化趋势。使用适应性矩估计(adaptive moment estimation,ADAM)算法优化模型使模型参数达到最优值,将均方误差(mean square error,MSE)作为损失函数,不断训练模型后最终得到其准确度为87.49%。使用支持向量机(support vector machine,SVM)模型预测新冠肺炎疫情发展的研究比较广泛,将其与LSTM模型进行对比,利用相同的数据集得出SVM模型预测的准确度为73.25%,对比2组数值发现,在预测印度新冠肺炎疫情上LSTM模型的准确度更高。该方法在一定程度上为预测印度新冠肺炎确诊病例数的研究提供了帮助,有助于人们实时监控印度疫情。

Since the outbreak of coronavirus in 2019,it has had a huge impact on the countries,societies and individuals.Some countries have taken some timely countermeasures,but the spread of the epidemic is still not fully under control.Machine learning algorithms are currently used to predict the development of the COVID-19 epidemic.The study uses LSTM model to analyze the data collected on cumulative confirmed cases in India to predict the trend of the epidemic in India.The model is optimized using the ADAM algorithm to optimize the model parameters,and the MSE is used as the loss function,and the model is trained continuously to obtain an accuracy of 87.49%.The use of SVM model to predict the development of COVID-19 epidemic is widely studied,and it is compared with LSTM model,and the accuracy of SVM model prediction is 73.25%using the same data set,and comparing the two values show that the LSTM model is more accurate in predicting the COVID-19 epidemic in India.This method has helped to predict the number of confirmed cases of COVID-19 in India to a certain extent and helps people to monitor the epidemic in India in real time.