Loading...
Publication
Modelação de dados espaço-temporais da COVID-19 em Portugal com desagregação temporal: uma abordagem estatística e epidemiológica
| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 8.67 MB | Adobe PDF |
Authors
Advisor(s)
Abstract(s)
A monitorização estatística analisa dados recolhidos ao longo do tempo e/ou espaço, para detetar e quantificar alterações, anomalias ou tendências nos processos subjacentes. Esta tese desenvolve e aplica métodos estatísticos para monitorizar dados espaço-temporais, considerando as suas especificidades e desafios.
Usando como caso prático a COVID-19 em concelhos de Portugal Continental, o trabalho visa compreender a dinâmica de disseminação de uma epidemia, apoiar a tomada de decisão em Saúde Pública e contribuir com metodologias inovadoras em estatística.
A tese propõe três contributos principais: 1) desagregação temporal de dados, em que se obtém dados mais detalhados e confiáveis a partir de dados de diferentes fontes que os recolhem e tratam com critérios distintos, recorrendo a algoritmos que permitem infundir informação relevante do domínio; 2) modelação de dados, considerando efeitos espaciais, temporais e de interação 𝑒𝑠𝑝𝑎ç𝑜 × 𝑡𝑒𝑚𝑝𝑜, com recurso ao ajuste de modelos hierárquicos bayesianos aos dados observados e desagregados, para análise e deteção de padrões temporais e espaciais; e 3) análise e deteção de fatores de risco e avaliação do impacto de eventos públicos e de covariáveis meteorológicas na disseminação de uma epidemia/pandemia, usando modelos lineares generalizados, visualização e testes de hipóteses. A tese explora dois casos relevantes com dados de COVID-19 de concelhos de Portugal Continental, criando formas mais sólidas de análise e monitorização dos dados.
A tese demonstra como a matemática e a epidemiologia se unem para enfrentar desafios da humanidade, como a pandemia COVID-19, avançando o conhecimento científico e melhorando as políticas públicas de saúde e outros processos.
Statistical monitoring analyses data collected over time and/or space to detect and quantify changes, anomalies, or trends in underlying processes. This thesis develops and applies statistical methods for monitoring spatiotemporal data, considering their specificities and challenges. Using COVID-19 in counties of Continental Portugal as a practical case, the work aims to understand the dynamics of epidemic spread, support decision-making in Public Health, and contribute with innovative statistical methodologies. The thesis proposes three main contributions: 1) the temporal disaggregation of data, which obtains more detailed and reliable data, combining data from different sources that collect and treat them with different criteria, using algorithms that allow infusing relevant information from the domain; 2) data modelling, considering spatial, temporal, and 𝑠𝑝𝑎𝑐𝑒 × 𝑡𝑖𝑚𝑒 interaction effects, using the adjustment of Hierarchical Bayesian Models to observed and disaggregated data for analysis and detection of temporal and spatial patterns; and 3) analysis and detection of risk factors and evaluation of the impact of public events and meteorological covariates on the spread of an epidemic/pandemic, using Generalized Linear Models, visualization, and hypothesis testing. The thesis explores two relevant cases with COVID-19 data from counties of Continental Portugal, creating more robust forms of analysis and monitoring of the data. The thesis demonstrates how mathematics and epidemiology come together to address humanity's challenges, such as the COVID-19 pandemic, advancing scientific knowledge and improving public health policies and other processes. The thesis demonstrates how mathematics and epidemiology come together to face challenges of humanity, such as the COVID-19 pandemic, advancing scientific knowledge and improving public health policies and other processes.
Statistical monitoring analyses data collected over time and/or space to detect and quantify changes, anomalies, or trends in underlying processes. This thesis develops and applies statistical methods for monitoring spatiotemporal data, considering their specificities and challenges. Using COVID-19 in counties of Continental Portugal as a practical case, the work aims to understand the dynamics of epidemic spread, support decision-making in Public Health, and contribute with innovative statistical methodologies. The thesis proposes three main contributions: 1) the temporal disaggregation of data, which obtains more detailed and reliable data, combining data from different sources that collect and treat them with different criteria, using algorithms that allow infusing relevant information from the domain; 2) data modelling, considering spatial, temporal, and 𝑠𝑝𝑎𝑐𝑒 × 𝑡𝑖𝑚𝑒 interaction effects, using the adjustment of Hierarchical Bayesian Models to observed and disaggregated data for analysis and detection of temporal and spatial patterns; and 3) analysis and detection of risk factors and evaluation of the impact of public events and meteorological covariates on the spread of an epidemic/pandemic, using Generalized Linear Models, visualization, and hypothesis testing. The thesis explores two relevant cases with COVID-19 data from counties of Continental Portugal, creating more robust forms of analysis and monitoring of the data. The thesis demonstrates how mathematics and epidemiology come together to address humanity's challenges, such as the COVID-19 pandemic, advancing scientific knowledge and improving public health policies and other processes. The thesis demonstrates how mathematics and epidemiology come together to face challenges of humanity, such as the COVID-19 pandemic, advancing scientific knowledge and improving public health policies and other processes.
Description
Tese de Doutoramento em Matemática Aplicada e Modelação apresentada à Universidade Aberta
Keywords
COVID-19 Dados espaciais Dados espaço-temporais Desagregação de dados temporais Epidemiologia Modelação de dados espaço-temporais Spatial data Spatiotemporal data Temporal data disaggregation Epidemiology Spatiotemporal data modelling
Citation
Leal, Maria da Conceição Dias - Modelação de dados espaço-temporais da COVID-19 em Portugal com desagregação temporal [Em linha]: uma abordagem estatística e epidemiológica. [S.l.]: [s.n.], 2024. 309 p.