Poster
Uplift Modeling from Separate Labels
Ikko Yamane · Florian Yger · Jamal Atif · Masashi Sugiyama

Thu Dec 6th 10:45 AM -- 12:45 PM @ Room 210 #5
in Thu Poster Session A »

Uplift modeling is aimed at estimating the incremental impact of an action on an individual's behavior, which is useful in various application domains such as targeted marketing (advertisement campaigns) and personalized medicine (medical treatments). Conventional methods of uplift modeling require every instance to be jointly equipped with two types of labels: the taken action and its outcome. However, obtaining two labels for each instance at the same time is difficult or expensive in many real-world problems. In this paper, we propose a novel method of uplift modeling that is applicable to a more practical setting where only one type of labels is available for each instance. We show a mean squared error bound for the proposed estimator and demonstrate its effectiveness through experiments.

Keywords: Semi-Supervised Learning · Causal Inference

Author Information

Ikko Yamane (The University of Tokyo/RIKEN)
Florian Yger (Université Paris-Dauphine)
Jamal Atif (Université Paris-Dauphine)
Masashi Sugiyama (RIKEN / University of Tokyo)

More from the Same Authors