Forecasting in SAP Lumira 1.25

The new 1.25 forecasting capability in SAP Lumira lets you use historical data as the basis for predicting future values.

The forecasting feature analyzes the trends and cycles of a time series to predict future values. Forecasting uses a measure and a dimension that is part of a time hierarchy (for example, Month) as its inputs. You specify how many forecasted values you want the algorithm to produce. SAP Lumira provides two algorithms for forecasting future data:

• SAP Predictive Analytics, and
• Triple Exponential Smoothing.

SAP Predictive Analytics Time Series Analysis computes several models that are compared for best results. It does this by breaking a time series into four components:

• Trend
• Cycles
• Fluctuations
• Information Residue

Forecasting Example

Here is an example of how SAP Predictive Analytics Time Series Analysis can predict future values. The blue line on the chart represents the actual data, and the red line represents the SAP Predictive Analytics Time Series Analysis forecast.

In addition to SAP Predictive Analytics forecasting, the Triple Exponential Smoothing (TES) algorithm can be used, in particular for situations where SAP Predictive Analytics Time Series Analysis can’t produce a forecast. The TES algorithm always produces a result, but the result is usually of lower quality.

Want to learn more about Forecasting?

To learn more about how forecasting and other predictive analytics techniques can help your business succeed, visit our SAP Predictive Analytics site: http://www.sap.com/learn-predictive.