In a recent development, on 1 million simulated customers, calibrated Qini scored a Linear DR learner 0.0856 against 0.0826 for the true-CATE ranking that should have won. The post When Qini Rewards the Wrong Uplift Model, Use Doubly Robust Scores appeared first on Towards Data Science.

On 1 million simulated customers, calibrated Qini scored a Linear DR learner 0.0856 against 0.0826 for the true-CATE ranking that should have won.

The post When Qini Rewards the Wrong Uplift Model, Use Doubly Robust Scores appeared first on Towards Data Science.

→ Read original article