Application of the Fuzzy Auto- Regressive Integrated Moving Average (FARIMA) Model for Iran s Steel Consumption Forecasting

The use of non-stochastic models such as fuzzy time seriesforecasting models for time series analysis has attracted theattention of researchers in recent years. Auto-RegressiveIntegrated Moving Average (ARIMA) models are one of themost important time series models used in financial marketforecasting. Recent research activities in time seriesforecasting indicate that two basic limitations detract fromtheir popularity for time series forecasting: (1) ARIMAmodels assume that future value of a time series have alinear relationship with current and past values as well aswith white noise. (2) ARIMA models require a large amountof historical data in order to produce accurate results.Fuzzy autoregressive integrated moving average (FARIMA)models are the fuzzy improved version of the autoregressiveintegrated moving average (ARIMA) models, proposed inorder to overcome limitations of the traditional ARIMAmodels; especially data limitation, and yield more accurateresults. Empirical results of Iran s steel consumptionforecasting indicate that the proposed model exhibiteffectively improved forecasting accuracy, so it can be usedas an alternative model to steel consumption forecasting,especially when the scrimpy data made available.

Keywords:

Auto- Regressive Integrated Moving Average , Fuzzy models , Time series forecasting , crude Steel consumption