352-361
Generating temporal model using climate variables for the prediction of dengue cases in Subang Jaya, Malaysia
Authors: Nazri Che Dom, A Abu Hassan, Z Abd Latif, Rodziah Ismail
Number of views: 300
Objective: To develop a forecasting model for the incidence of dengue cases in Subang Jaya
using time series analysis.
Methods: The model was performed using the Autoregressive Integrated Moving Average (ARIMA)
based on data collected from 2005 to 2010. The fitted model was then used to predict dengue
incidence for the year 2010 by extrapolating dengue patterns using three different approaches
(i.e. 52, 13 and 4 weeks ahead). Finally cross correlation between dengue incidence and climate
variable was computed over a range of lags in order to identify significant variables to be included
as external regressor.
Results: The result of this study revealed that the ARIMA (2,0,0) (0,0,1)52 model developed, closely
described the trends of dengue incidence and confirmed the existence of dengue fever cases in
Subang Jaya for the year 2005 to 2010. The prediction per period of 4 weeks ahead for ARIMA (2,0,0)
(0,0,1)52 was found to be best fit and consistent with the observed dengue incidence based on the
training data from 2005 to 2010 (Root Mean Square Error=0.61). The predictive power of ARIMA (2,0,0)
(0,0,1)52 is enhanced by the inclusion of climate variables as external regressor to forecast the
dengue cases for the year 2010.
Conclusions: The ARIMA model with weekly variation is a useful tool for disease control and
prevention program as it is able to effectively predict the number of dengue cases in Malaysia.