[ad_1]
Fixed sensors provide data on pedestrian traffic in metropolitan areas. Mobile sensors provide information based on anonymous mobile phone data from applications such as Google Maps. WTP data provides information on coronavirus particles in wastewater. Weather data provides information on temperature and humidity. Policy data provides information on the day-to-day policy measures implemented by governments in response to the pandemic.
The relationship between the indicator and weekly COVID-19 hospitalizations is estimated using a structural time series (STS) model. The STS model decomposes the observed hospitalizations into trend and regression components. Each time-dependent regression coefficient describes how the relationship between the auxiliary series and hospitalizations evolves during the COVID-19 pandemic. The trend components considered are local level, smooth trend, and local linear trend models as well as models with time-invariant intercepts. The relevant auxiliary variables for the regression components are selected using a stepwise variable selection method. After expressing the model in state space form, the model is fitted using a Kalman filter.
Different model results were extensively evaluated and discussed. Therefore, this article aims to provide a comprehensive comparison of the various models, their results, and their implications, in addition to presenting the potential relevance of indicators for predicting COVID-19 hospitalizations.
Our analysis revealed significant relationships between various indicators and COVID-19 hospitalizations, highlighting the importance of specific indicators for modeling and forecasting hospitalizations during the pandemic. By comprehensively evaluating the predictive power of different indicators and model specifications, we facilitate informed decision-making during the pandemic, thereby promoting a more effective public health response.
[ad_2]
Source link
