Dengue em Natal/RN: uma análise do período 2000-2016 via séries temporais

Abstract

Dengue fever is an infectious disease transmitted by the Aedes aegypti mosquito. This vector also transmits the chikungunya, zika and yellow fever. In 2019, the World Health Organization (WHO) has established the fight against dengue as one of the ten priorities for this year. It is estimated that almost half of the world’s population is at risk of dengue infection. Based on the above, this study aims to analyze data on reported cases of dengue fever between 2000 and 2016, obtained from the Center for Zoonoses Control (CCZ) in the city of Natal/RN, through time series. In Natal/RN since 2000 dengue outbreaks occur, implying the existence of possible atypical observations (denoted in this paper by outliers) in the historical series. In addition, an outbreak of zika occurred in 2015 and because of a similar symptoms, cases may have been reported as dengue. Therefore, we search for a model that considers information of possible pattern changes and existence of outliers by intervention analysis. The exponential smoothing methods simple, Holt and HoltWinters were also used, as well as the ARIMAX model with exogenous climatic variables. Models are compared using root mean square error (rMSE) and mean absolute error (MAE) of forecasts for the first 37 weeks of 2017. The GARCH model was used to estimate the volatility of the series.