Combined Forecasting Techniques for Epidemiological Surveillance Trend

摘要

This research aims to study combined forecasting techniques by weighted based on Minimum Mean Absolute Error Method (MMAE) with goal programming. There are three classical individual forecast techniques; classical decomposition, exponential smoothing, Box-Jenkins techniques and combined forecasts models are studied and compared. Secondary data are used to forecast the patient numbers quarterly with diagnosis of dengue hemorrhagic fever (DHF) in Thailand for the period of 2003-2011, out of which data till December 2012 were used to check the forecasting ability of the model. Results during this period confirm that Minimum Mean Absolute Error Method (MMAE) provides a better forecasting of patient numbers with DHF than others models. Efficiency comparison of forecasting models by using MAPE shows that the proposed method is the best among all these models.