Predictive Forecasting Through Integrated Planning Models Smart Predict uses both planning models and data sets for building time series forecasting models, relying on separate sources for training and predictions. Distinct planning model versions store historical data and forecast outcomes, ensuring that technical compatibility is maintained between acquired or live datasets in a single SAP HANA instance. This design integrates predictive planning to harness the full capabilities of the underlying planning models, ensuring accurate forecast outputs.
Structured Data Sets for Precise Forecasting Accurate forecasting depends on data sets with a strict structure where each observation has a unique date and a measured value formatted properly. Explicit forecast horizon dates can be provided to enforce granularity when automatic date inference might be skewed by missing data. Supplementing the data with influencer columns captures external factors, while up to five entity columns allow segmentation for targeted predictions. Ensuring that every forecast period has complete, non-aggregated data is essential for reliable prediction results.