In: Operations Management
Discuss the similarities and difference between Level Shifts and Additive Outliers including how you would go about identifying them and then dealing with them in a forecasting scenario
The similarities between Additive Outlier and Level Shift Outlier are as follows:
· Both are types of outliers i.e. are inconsistent when compared with the majority of the observation
· Both have considerable influence on the forecasting accuracy of the time series model.
The fundamental differences between Additive Outlier and Level Shift Outlier are as follows:
· Additive Outlier appears as an isolated spike in the graph while Level Shift outlier can depict a seasonal or abrupt change in the course of the graph
· The subsequent observations before or after the spike of additive outlier remain unaffected by its occurrence. Level Shift Outlier has a considerable impact on the observations which happen after its occurrence.
· Additive Outlier can be considered to have a temporary effect on the forecasting method while a Level Shift Outlier tends to gave a permanent effect on the forecasting model.