Measuring the impact of AI

The boost in the success rate, also known as lift, that artificial intelligence (AI) achieves is an important business metric. To report on this metric, a data scientist can use predictions. Predictions add best practices to predictive models and use a control group as a benchmark to measure lift. Customers in the control group receive a random offer instead of the one selected by AI. The use of a control group also adds a degree of exploration to the exploitation of the models.

Débutant

Anglais

Après avoir terminé ce module, vous pourrez :

Describe how predictions add best practices to predictive models

Explain how the use of a model control group allows the measurement of lift

Explain how the use of a model control group adds exploration to the exploitation of the models

Rubriques

Measuring lift using a control group

10 mins

Quiz du module

5 mins