OBJECTIVES: The target of the Fourth Millennium Development Goal (MDG-4) is to reduce the rate of under-five mortality by two-thirds between 1990 and 2015. a periodicity of six months in the fitted model are very likely to be consequences of poor quality of data collection and reporting systems. CONCLUSIONS: The present work is the first attempt at time series modeling of the U5MR in Iran, and reveals that improvement of under-five mortality data collection in health facilities and their corresponding systems is a major challenge Serpinf1 to fully achieving the MGD-4 in Iran. Studies similar to the present work can enhance the understanding of the invisible patterns in U5MR, monitor progress towards the MGD-4, and predict the impact of future variations on the U5MR. is a white noise process with standard deviation to be stationary but it assumes that after d times differencing, the new series becomes a stationary zero mean time series. It also assumes that the current value of the differenced series is a linear combination of previous values and of a white noise process (independent and identically distributed random shocks with zero mean and common variance is the number of terms in the model that describe the direct dependency among successive Torcetrapib observations and the MA order is the number of persistent random shocks that describe indirect dependency among successive observations [25]. They are identified by examining patterns in plots of the autocorrelation function (ACF) and partial autocorrelation function (PACF) of the differenced series [23,25]. After identifying and and and seem plausible, model selection criteria such as the Akaike information criterion (AIC) and the Schwarzs Bayesian information criterion (BIC) can be used to select the best fitting model [24]. Seasonal or periodic cycles produce local spikes at specific lags in plots of ACF and PACF [23,25]. Any ARIMA model can be extended to include seasonality and is then called a seasonal autoregressive integrated moving average (SARIMA) model. A SARIMA (and difference order [17,26,27]. The seasonal component is responsible for the recurrence of a recognizable pattern after the seasonal periods. The general model equation of a SARIMA (regular autoregressive parameters, regular MA parameters, seasonal autoregressive parameters and seasonal MA parameters [17]. Similar to the regular ARIMA model, ACF and PACF plots and AIC and BIC criteria can be used to identify Torcetrapib the seasonal orders (which implies that the model residuals be normally distributed [17]. Plots of ACF and PACF of the model residuals are useful in assessing no significant serial correlation in the model residuals. If the fitted model is adequate, both the ACF and PACF of residuals should have no structure to identify [24,25]. In addition, the portmanteau Q-test (goodness-of-fit statistic) and Bartletts periodogram-based test for white noise can be also used to test the existence of serial correlation in the model residuals [23]. The adequacy of the normality assumption of the model residuals can be checked using normality tests such as the Shapiro-Wilk, and Skewness-Kurtosis tests [25,28]. Finally, the heteroscedasticity (deviation from constant variance assumption) in the model residuals can be examined by performing an ARCH-LM test for autoregressive conditional heteroscedasticity effects [23,24]. Prediction and prediction intervals Once a suitable model has been fitted to the data, the known Torcetrapib structure of the model can be used to predict future values of the time series based on previously observed values with a recursive method [23]. Prediction error, the difference between the actual and the predicted values, is inevitable and hence the uncertainty of the predictions must be considered; the the prediction beyond the noticed beliefs further, the less dependable the prediction. For this good reason, an period that with a particular.