Time series prediction of under-five mortality rates for Nigeria: comparative analysis of artificial neural networks, Holt-Winters exponential smoothing and autoregressive integrated moving average models

Table 1 Performance measures of time series techniques for under-five mortality rates in Nigeria

Model	GMDH-type ANN	ARIMA	Holt-Winters exponential smoothing
Best parameters	Training set = 80%, testing set = 20%	p = 0, d = 3, q = 2	α=0.91, β=0.51
RMSE	0.09	0.23	2.87
RMAE	0.25	0.41	1.20
Modified NSE	0.998	0.996	0.967
DM test statistic	Reference	−3.608*	−4.474*

^*significant at p-value < 0.05; p = number of autoregressive terms, d = number of differencing, q = number of moving average terms; RMSE: Root mean squared error; RMAE: Root mean absolute error; α=coefficient for the level smoothing; β= coefficient for the trend smoothing; modified NSE: modified Nash-Sutcliffe model efficiency coefficient; DM: Diebold-Mariano test

ISSN: 1471-2288