Abstract
This paper presents an iterated liner regression model and compares its forecasting performance with the traditional liner regression (LR) and BoxJenkins ARIMA models using two well-known time series datasets: airline data and sunspot data. The difference between iterated LR and traditional LR is the former considers the error and uses it as dependent variables again to reduce the error rate until error rate is very small. The results show that the performance of iterated LR is slightly better than Box-J model and much better than traditional LR models.