Small Improvements Causing Substantial Savings - Forecasting Intermittent Demand Data Using SAS Forecast Server (2008)

by Michael Leonard. Bruce Elsheimer. Meredith John. Udo Sglavo.

Businesses require accurate forecasts of time series data that is not continuous. Often, time series data is intermittent (discontinuous or interrupted). Intermittent time series data points are mostly zero (the base value), with occasional departures from the base value. Intermittent time series are common in business and economic data. For example, at progressively lower levels of data disaggregation (larger frequency, smaller geography, or both), the time series data is often intermittent. The most commonly used forecasting techniques are continuous time series methods such as exponential smoothing methods (ESM). Continuous methods are meant to forecast the future values with respect to future time periods. Because the most likely future value is zero (the base value), these models are inadequate when compared to the naïve model of simply zero (the base value). In contrast, intermittent demand methods (IDM) forecast the future average demand per period, which is more appropriate, especially for many inventory control systems. Additionally, IDM is useful for forecasting time series data that is hierarchical (for example, when the upper levels of aggregation are continuous and the lower levels of disaggregation are intermittent). This paper exposes the inadequacy of continuous time series methods when compared to IDM for forecasting future average demand per period for intermittent time series. This paper demonstrates a technique and system of large-scale automatic forecasting of intermittent demand series. This paper explains how SAS Forecast Server is used as this system. [via]

PDF Finder

Small Improvements Causing Substantial Savings - Forecasting Intermittent Demand Data Using SAS Forecast Server (2008)

Related Files

Rate this Document

Popular eBooks

ADS

Tag Clouds

Last Download

BookShelf