Predição semanal de números de casos de dengue em Porto Alegre por modelos de séries temporais

Abstract

This Work develops and evaluates time series models for the weekly prediction of dengue cases in the city of Porto Alegre, with the aim of offering effective tools for epidemiological surveillance and planning preventive measures. Using public health data and region-specific meteorological records, statistical models such as SARIMA (Seasonal Autoregressive Integrated Moving Average) and SARIMAX (Seasonal Autoregressive Integrated Moving Average with Exogenous Variables) were applied and compared in terms of their predictive efficiency. The study highlighted that the inclusion of climatic variables, such as temperature and precipitation, in the SARIMAX models contributes to increasing the accuracy of predictions, adapting to dengue transmission dynamics, which are strongly influenced by environmental conditions. The work also emphasizes the importance of time series analysis in public health, especially for diseases with a strong seasonal component and dependence on weather conditions, such as dengue.