TREND ANALYSIS AND FORECASTING MODELS FOR UNDER-FIVE MORTALITY RATE IN MALAYSIA

Abstract

The aim of this research is to study the trend pattern, develop forecasting models and forecast
future trends of under-five mortality (U5MR) in Malaysia by gender. The Box-Jenkins Model
and State Space Model are used in forecasting the yearly U5MR of a 37-year period (1980 –
2016) in Malaysia. Random Walk model for Box Jenkins and Local Linear model for State
Space model are identified as the best models to suit the U5MR data characteristics. The
performance of these two models are evaluated based on in-sample and out-sample
evaluation using lowest root mean square error (RMSE) and mean absolute percentage error
(MAPE). The results show that U5MR in Malaysia fluctuated from year to year with a slowly
decreasing trend pattern for both genders with male recording a higher rate compared to
female. Moreover, the Local Linear Trend model is chosen as the best model for forecasting
future U5MR data with the lowest value of RMSE and MAPE. Future trend slightly increased
from 0.00189 from 0.001891 in 2017 to 0.001907 in 2030 for male while for female is
0.001667 in year 2017 to 0.001878 in 2030. Thus, it shows the forecast trend of Malaysia’s
U5MR for the male population is higher than the female. This study is significant in that it
may become a reference for other countries as an indicator for human resource management
and health care allocation planning