Using State Space Methods to Detect When a Forecast Begins to Go Out of Control
12 June 1988
It is important to know early on when a forecast begins to deviate from the actuals. When it is the locus of a life cycle curve which is being forecasted, information about such deviations becomes vital because it may signal that a life cycle turning point is in the making or that the forecast will have to be recalibrated. This paper develops a state space model which is capable of making such detections. Key words.: Forecasting; Statistical process control; Kalman filter; Adaptive tracking and control.