Forecasting models for Dengue outbreak in Northern province of Sri Lanka

Abstract

The Dengue Hemorrhagic fever is one of frequently chronicled epidemiological scenario in thepast and ongoing time frame worldwide. Recently, trend of its outbreak seems numerousincidents partially deviated compare to existing pattern from the past literature. Institutionalbodies involved in pandemic alleviation of Dengue outbreak in great difficulties to develop anefficient management plan because of complicated pattern of its outbreak. This study aimedtowards finding an appropriate time series forecasting model for the number of Dengue incidentsrecorded in Northern province of Sri Lanka by tracing of Auto Regressive Integrated MovingAverage (ARIMA) model regardless of climatic factors such as, ambient temperature, rainfall andalso the demographic factors associated with the data set for the period of January 2009 toDecember 2014 on monthly basis. Appropriate model was derived through sequential processincludes; formulate stationary series, tentative model selection, model optimization, modelvalidation and error & accuracy analysis. As far as the appropriateness was concerned throughoutabove processes concluded that the best fitted model for the Northern province isARIMA (3, 1, 4) model which could applicable to forecast the dengue incidences in Northernprovince for near future and it could be more reinforced by considering weather factors in future.Morbidity of Dengue in Northern province has been recorded as in intermittently increasingduring the study period while regular difference data series as stationary. At the present timebeing ARIMA (3, 1, 4) model could applicable for the successful alleviation of Dengueconsequently shifting towards socio-economic advancement of the nation.