|Saturday, February 23|
|CS16 Theme 2: Data Modeling and Analysis #6||
Sat, Feb 23, 10:45 AM - 12:15 PM
Estimation of Individual Treatment Effects (302434)
Keywords: causal effects, personalized medicine, direct marketing, model selection, uplift modeling
Estimating the effect of a treatment at the individual level is of tremendous importance in personalized medicine, direct marketing, public policy and other areas. When treatment effects are heterogeneous within a population and treatments can be assigned individually, knowledge of individual treatment effects can be used to maximize the mean effectiveness of a treatment program. However, estimation of individual treatment effects is challenging because each individual in an experiment is typically in either the treatment group or the control group, so individual treatment effects are not observed. This talk will provide a survey of the current research in this area, summarizing and comparing standard solutions to several methods that have recently been developed. The model selection problem; that is, the problem of choosing between competing estimators of individual treatment effects, will also be discussed.