Forecasting Dengue Fever Epidemics using ARIMA Model

Abstract

Despite the control and surveillance of dengue fever in various ways, but still found many dengue patients in Thailand every year. The main objective of the research is to develop a model for the prediction of dengue fever. Model development on data of dengue fever outbreaks in Northeast Thailand. The ARIMA model and the Gaussian Distribution model to predict the incidence of dengue fever. Dengue patient data from 2015 to 2019 are used to validate the accuracy of predictive models. The autocorrelation function is measured as the correlation between dengue fever data of lag suggested parameters ARIMA (2,1,2), and partial autocorrelation function is defined as the difference between the autocorrelation coefficient at a certain lag 20 with ARIMA (1,2,2). The best models are ARIMA (2,1,2) evaluated a great forecast dengue fever epidemic with a lower mean absolute percentage error (MAPE) of 1148.319, lower Bayesian information criterion (BIC) of 13679.5133 and Akaike information criterion (AIC) of 13710.0388. The results will benefit health professionals. Moreover, the model can be used for policymaking and planning of resource allocation for the people and continue to improve public health services.