Time series forecasting using improved ARIMA

In recent years, there has been an explosion of interest in forecasting time series databases in different applied areas. Forecasting is one of the main goal's mining of time series databases. Time series forecasting has been shown effective in suitable decision making in various domains. So far, a variety of techniques have been proposed to obtain goal of prediction and analysis of literature this area is in different directions. In this regard, in this paper, there are two goals. First, provide a review. For this goal, this paper classifies previous major works that investigated the forecasting of time series data in different application areas. Second, propose a novel approach to improve ARIMA model by applying a mean of estimation error for time series forecasting. Experimental results indicate that the proposed approach can improve performance in the process of time series data forecasting.