Modeling and Predicting of Iran's Economic Growth Using ANFIS, Markov Switching and ARIMA Models

Economic growth forecast is a major problem in economy that has a significant impact in government policy and economic planning. It also helps policy makers for future decisions. Multivariate econometric forecasting models associated with many limitations, so an alternative approach is the use of univariate models, but most of these methods need a lot of data to achieve the best result. In this study, data from 1959 to 2005 were used to estimate the models. Then the performances of auto regressive integrated moving average (ARIMA) model in the economic growth forecast of Iran was evaluated and compared with Markov switching method and fuzzy neural network (ANFIS) for the period from 2006 to 2013 using the Criteria RMSE, MAE and MAPE. Results showed that ANFIS model had the best performance. Furthermore, Markov switching method was more suitable than ARIMA model.