You are here:
Explore Predictions and Improvements
Explore predicted outcomes and suggested ways to improve those predicted outcomes. Perform interactive, “what if” analyses and change feature selections to see prediction scores, top prediction factors, and top improvements to enhance prediction scores.
About Predictions and Improvements
In Einstein Discovery, a prediction is a probabilistic value of a future outcome, as calculated by a predictive model. Predictions are similar to the kind of regression or machine-learning analysis that data scientists conduct using advanced analytics tools. Knowing the most influential predictor variables can, for example, help a sales or marketing manager choose where to invest budget and which areas are less important. You can explore which variable values maximize or minimize your desired outcome.
An improvement is a suggested action, based on prescriptive analytics, that a user can take to improve the likelihood of a desired outcome. Improvements are associated with actionable variables, which are explanatory variables that people can control. For example, you can determine which marketing campaign or sales strategy to use for a given customer. Einstein Discovery recommends specific actions for changing predictors and seeing the impact on the prediction.
Explore What Could Happen
To simulate and explore future outcomes:
- Open a Model.
-
Click Predictions (1).

-
Under Select a Group to Predict (2), select an option from one or more features.

Field Description Einstein Prediction Prediction score for your selections. Probabilistic calculation of the outcome based on the model. Top Improvements Suggested actions that you can take to improve the predicted outcome. Model Overview Metrics that describe the quality of the model. For details, see Evaluate Model Quality. Top Prediction Factors Explanatory variables, favorable and unfavorable, that are most strongly associated with the predicted outcome. Insights Any insights associated with your prediction.
Note A shield icon indicates that the variable is flagged as sensitive. -
To see recommendations for a given feature, select
Actionable.

If there are ways to statistically improve the predicted outcome for the selected feature, Einstein displays them in Top Improvements.

