Previsão da Série Temporal de temperatura média da cidade de Cuiabá por meio de modelos Holt-Winters do tipo Aditivo e Multiplicativo com parâmetros ajustados via PSO

Abstract

Global warming has triggered a series of climate changes on a global scale, predominantly due to the gradual increase in temperature over the years. Therefore, it becomes essential for society to accurately monitor temperature increase trends in urban areas and develop effective public policies to reduce the impact of rising temperatures on the spread of vector-borne diseases and the health challenges resulting from prolonged exposure to high temperatures. Furthermore, this analysis provides a solid foundation for the development of urban heat mitigation policies, which, in turn, significantly contribute to improving the quality of life for urban residents. The scientific community has conducted a variety of studies aimed at understanding climate change. These investigations include specific methodologies designed to monitor climate elements such as air humidity and assess their impact on daily temperature variation. Among the approaches explored, the Box-Jenkins methodology is notable for its application in time series. In this context, this work aimed to conduct a time series forecasting experiment by applying the Particle Swarm Optimization (PSO) algorithm to adjust Holt-Winters models, encompassing both the multiplicative and additive approaches. The time series data used consisted of the daily average temperature recorded in the city of Cuiabá, located in the state of Mato Grosso. In the methodology employed, both the multiplicative and additive approaches were considered. Furthermore, the PSO algorithm was used to optimize the model parameters with the goal of obtaining accurate predictions.