Predicting COVID-19 Using Hybrid AI Model
17 Pages Posted: 24 Mar 2020More...
Background: The coronavirus disease 2019 (COVID-19) breaking out in late December 2019 is gradually being controlled in China, but it is still spreading in other countries and regions worldwide. It is urgent to conduct prediction research on the development and spread of the epidemic.
Methods: A hybrid AI model is proposed for COVID-19 prediction. First, by analyzing the change in the infectious capacity of virus carriers within a few days after infection, an improved SI (ISI) model is proposed. Second, considering the effects of prevention and control measures and the increase of the public’s prevention awareness, the natural language processing (NLP) module and the long short-term memory (LSTM) network are embedded into the ISI model to build the hybrid AI model for COVID-19 prediction.
Findings: Compared with traditional epidemic models, the proposed ISI model finds that the new confirmed cases of COVID-19 are mainly infected by the cases from previous 3 to 8 days, and the average infection time is about 5.5 days. With introduction of NLP and LSTM into the hybrid AI model, the mean absolute percentage errors (MAPE) of the prediction results are 0.52%, 0.38%, 0.05%, 0.86% for the next 6 days in Wuhan, Beijing, Shanghai and nationwide respectively, which proves our model is more in line with the actual epidemic development trend.
Interpretation: The infectious capacity of virus carriers varies at different stages. Additionally, both the prevention and control measures and the public’s awareness of the epidemic have great impact on the transmission of COVID-19.
Funding Statement: This study was supported by the fund from the National Key Research and Development Program of China (Grant No. 2016YFB1000900).
Declaration of Interests: The authors declare that they have no competing interests.
Suggested Citation: Suggested Citation