ARIMA forecasting of COVID-19 incidence in Italy, Russia, and the USA

Gaetano Perone

DOI: https://doi.org/10.48550/arXiv.2006.01754

2020-06-05

Abstract:The novel Coronavirus disease (COVID-19) is a severe respiratory infection that officially occurred in Wuhan, China, in December 2019. In late February, the disease began to spread quickly across the world, causing serious health, social, and economic emergencies. This paper aims to forecast the short to medium-term incidence of COVID-19 epidemic through the medium of an autoregressive integrated moving average (ARIMA) model, applied to Italy, Russia, and the USA The analysis is carried out on the number of new daily confirmed COVID-19 cases, collected by Worldometer website. The best ARIMA models are Italy (4,2,4), Russia (1,2,1), and the USA (6,2,3). The results show that: i) ARIMA models are reliable enough when new daily cases begin to stabilize; ii) Italy, the USA, and Russia reached the peak of COVID-19 infections in mid-April, mid-May, and late May, respectively; and iii) Russia and the USA will require much more time than Italy to drop COVID-19 cases near zero. This may suggest the importance of the application of quick and effective lockdown measures, which have been relatively stricter in Italy. Therefore, even if the results should be interpreted with caution, ARIMA models seem to be a good tool that can help the health authorities to monitor the diffusion of the outbreak.

Applications

What problem does this paper attempt to address?

The problem that this paper attempts to solve is to predict the number of newly confirmed cases of COVID - 19 in Italy, Russia and the United States in the short - to - medium - term by using the Autoregressive Integrated Moving Average (ARIMA) model. Specifically, the author hopes:

1. **Predict the epidemic trend**: Predict the daily number of newly confirmed cases in these countries in the next 30 days through the ARIMA model.
2. **Determine the inflection point of the epidemic**: Identify the peak of the epidemic in each country and predict when the daily number of new cases will drop to near zero.
3. **Evaluate the impact of lockdown measures**: By comparing the prediction results of different countries, explore the importance of strict lockdown measures in controlling the spread of the epidemic.

### Key issues
- **Peak of the epidemic**: Italy, the United States and Russia reached the peak of the epidemic in mid - April, mid - May and late May respectively.
- **Time for the epidemic to subside**: The prediction results show that Russia and the United States will need more time than Italy to make the daily number of new cases drop to near zero.
- **Effect of lockdown measures**: Strict lockdown measures (such as those implemented in Italy) are helpful in controlling the epidemic more quickly.

### Methodology
The author used the non - seasonal ARIMA model to analyze the time - series data of the daily number of newly confirmed cases. The selection of the model was based on the following criteria:
- **AICc (Akaike Information Criterion corrected)** and **Maximum Likelihood Estimation (MLE)** were used to determine the best ARIMA parameters.
- **Prediction error measurement**: Including Root Mean Square Error (RMSE), Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE) and Mean Absolute Scaled Error (MASE) to evaluate the accuracy of the model.
- **Residual analysis**: Verify the assumptions of the model (such as the independence and homoscedasticity of residuals) through the Ljung - Box test and Engle's LM test.

### Results
- **Italy**: The ARIMA(4, 2, 4) model performed best, and the prediction results showed that the daily number of new cases would approach zero in mid - June.
- **Russia**: The ARIMA(1, 2, 1) model performed relatively well, predicting that the daily number of new cases would approach zero at the end of August.
- **USA**: The ARIMA(6, 2, 3) model performed best, predicting that the daily number of new cases would approach zero from the end of September to the beginning of October.

### Conclusion
The ARIMA model can be used as a simple and effective tool to help health departments monitor the development of the epidemic and allocate resources rationally. However, the author also pointed out that although the ARIMA model showed high reliability in short - and medium - term predictions, the results still need to be interpreted with caution, and it is recommended to continuously update data to improve the prediction accuracy.