A new adaptive exponential smoothing method for non-stationary time series with level shifts

Journal of Industrial Engineering International

دوره 10، شماره 4، اسفند 2014   

نویسندگان

Mohammad Ali Saniee Monfared^{* 1}؛ Razieh Ghandali²؛ Maryam Esmaeili¹

¹School of Engineering, Alzahra University, 1993891176, Tehran, Iran

²Industrial Engineering Department, Yazd University of Technology, Yazd, Iran

چکیده

Simple exponential smoothing (SES) methods are the most commonly used methods in forecasting and time series analysis. However, they are generally insensitive to non-stationary structural events such as level shifts, ramp shifts, and spikes or impulses. Similar to that of outliers in stationary time series, these non-stationary events will lead to increased level of errors in the forecasting process. This paper generalizes the SES method into a new adaptive method called revised simple exponential smoothing (RSES), as an alternative method to recognize non-stationary level shifts in the time series. We show that the new method improves the accuracy of the forecasting process. This is done by controlling the number of observations and the smoothing parameter in an adaptive approach, and in accordance with the laws of statistical control limits and the Bayes rule of conditioning. We use a numerical example to show how the new RSES method outperforms its traditional counterpart, SES.

کلیدواژه‌ها

Time series analysis . Adaptive exponential smoothing . Level shifts . Statistical control limits