Outliers

Outliers are extreme demand points in historical demand that you can't explain using the forecasting process.

The forecasting engine removes historical demand data, such as shipments and bookings, to detect outliers, and replaces the missing data to minimize forecast error. In such instances, the forecasting methods can adjust historical demand and smoothen extreme data points. The forecasting process also marks the points which were smoothed.

Understanding where outliers occurred can help in forecast analysis. You can do this using various reports and mark outliers for review in tables and graphs.

Detailed plan outlier reports display history and smoothed outliers. The reports enable you to easily identify the following:

• The percentage of forecast combinations having outliers.

• The number of combinations where outliers are detected and smoothed.

• Average number of outliers found in a combination.

Predefined layouts that you can use for viewing outliers are the following:

• Bookings Forecast Outlier Analysis

• Bookings Forecast Outlier Details

• Shipments Forecast Outlier Analysis

• Shipments Forecast Outlier Details

In addition, you can add the measures showing outlier values to any table or graph that shows historical demand. This can help you understand how historical demand was modified during the forecast.