Zhang (2003) proposed a hybrid ARIMA-ANN model, which outperformed the individual models for Wolf's sunspot data, Canadian lynx data and exchange rate time series data. Another hybrid SARIMA and ANN was presented by Aburto and Weber (2007) to forecast daily demand in a Chilean supermarket. Their proposed model outperforms existing naïve, seasonal naïve, unconditional average, SARIMAX and several neural network models. Based these studies, many of them tried to apply hybrid ARIMA-ANN model in different fields of application. Maia et al. (2008) presented an interval-valued time series forecasting of meteorological and stock price data using a hybrid ARIMA-ANN model and compared it with autoregression (AR), ARIMA and ANN models. Díaz-Robles et al. (2008) introduced a hybrid ARIMA-ANN model to forecast air quality in Chile. Cadenas and Rivera (2010) recommended a hybrid ARIMA-ANN model to forecast wind speed. Shukur and Lee (2015) also suggested the use of hybrid kalman filter and artificial neural network (KF-ANN) model to improve the accuracy of wind speed forecasting. Jeong et al. (2014) proposed a hybrid SARIMA-ANN model, which was shown to provide better accuracy for forecasting annual energy cost budget (AECB) in educational facilities of South Korea than the classical SARIMA model. Babu and Reddy (2014) suggested a hybrid ARIMA-ANN model, which has higher accuracy when compared to the individual models and the existing hybrid ARIMA-ANN models using a simulated data set and experimental data sets such as sunspot data, electricity price data and stock market data.