Improving prediction of COVID-19 evolution by fusing epidemiological and mobility data

Abstract

We are witnessing the dramatic consequences of the COVID-19 pandemic which, unfortunately, go beyond the impact on the health system. Until herd immunity is achieved with vaccines, the only available mechanisms for controlling the pandemic are quarantines, perimeter closures and social distancing with the aim of reducing mobility. Governments only apply these measures for a reduced period, since they involve the closure of economic activities such as tourism, cultural activities, or nightlife. The main criterion for establishing these measures and planning socioeconomic subsidies is the evolution of infections. However, the collapse of the health system and the unpredictability of human behavior, among others, make it difficult to predict this evolution in the short to medium term. This article evaluates different models for the early prediction of the evolution of the COVID-19 pandemic to create a decision support system for policy-makers. We consider a wide branch of models including artificial neural networks such as LSTM and GRU and statistically based models such as autoregressive (AR) or ARIMA. Moreover, several consensus strategies to ensemble all models into one system are proposed to obtain better results in this uncertain environment. Finally, a multivariate model that includes mobility data provided by Google is proposed to better forecast trend changes in the 14-day CI. A real case study in Spain is evaluated, providing very accurate results for the prediction of 14-day CI in scenarios with and without trend changes, reaching 0.93 $$R^2$$ R 2 , 4.16 RMSE and 1.08 MAE.

COVIDSensing: Social Sensing Strategy for the Management of the COVID-19 Crisis

“…After running both techniques, the coefficient of determination (R 2 ) and root mean square error (RMSE) metrics are calculated, the predictions for the best values are jointly selected, and they are finally displayed to the user. We refer the reader to [45] for insights on the evaluation of these procedures. Fitting the ARIMA model is not straightforward, as the best combination of parameters has to be found to minimise the RMSE.…”

Using a stochastic continuous-time Markov chain model to examine alternative timing and duration of the COVID-19 lockdown in Kuwait: what can be done now?

“…Contact tracing implemented earlier in the epidemic in several countries provide better insights into the epidemic peak, which would help in gauging the perfect timing for implementation of a lockdown. For example, contact tracing can help in predicting the evolution of the COVID-19 infections so that predictions of the peak of the epidemic becomes easier [ 25 ].…”

Autoregressive count data modeling on mobility patterns to predict cases of COVID-19 infection

“…While the smart devices are GPS-tracked, the locations of the transaction data can be found at retail outlets, leisure facilities and other public amenities. Among the public datasets currently available, the Google COVID-19 Community Mobility Reports have been widely used for forecasting cases of infection and providing insights on how to use mobility characteristics efficiently [30][31][32][33][34]. Sufficient mobility records in both spatial and temporal dimensions enable the training of machine learning models that require large amount of data.…”

“…Sufficient mobility records in both spatial and temporal dimensions enable the training of machine learning models that require large amount of data. Some research has adopted a number of statistical and machine learning models based on a recurrent neural network as well as an ensemble approach in order to predict trend changes in the 14-day cumulative incidence [34]. In this work, two datasets with similar training periods but different testing periods were used to compare the models.…”