SAC Predictive Scenarios makes machine learning accessible to business analysts — building classification models for customer churn without advanced technical expertise.
SAP Analytics Cloud (SAC) Predictive Scenarios simplifies machine-learning implementation through user-friendly tools that empower business analysts to generate meaningful insight from data — eliminating the need for manual algorithm development and making predictive analytics accessible to non-technical users.
Three core scenario types
The tool offers classifications, regressions and time-series forecasts. This focuses on classification, which sorts data into distinct categories — in banking, separating customers likely to churn from those likely to remain, analysing binary variables (such as 'Exited' status) and calculating influencer contributions.
Implementation steps
Data import and preparation; statistical type definition (nominal, ordinal, continuous, textual); outlier detection and handling; model creation and training; and test-data application — with the platform automatically optimising influencer selection during training.
Model performance metrics
Predictive Power (precision; closer to 100% is better), Prediction Confidence (consistency on similar new datasets), and the Gini Index (0–1, assessing classification distribution) — together giving confidence in the model before applying it to support decisions.
