A0401
Title: Casual inference of general treatment effects using neural networks with a diverging number of confounders
Authors: Zheng Zhang - Renmin University of China (China) [presenting]
Abstract: Estimating causal effects is a primary goal of behavioural, social, economic and biomedical sciences. Under the unconfoundedness condition, adjustment for confounders requires estimating the nuisance functions relating outcome and/or treatment of confounders. A generalized optimization framework is considered for efficient estimation of general treatment effects using feedforward artificial neural networks (ANNs) when the number of covariates is allowed to increase with the sample size. ANNs estimate the nuisance function, and a new approximation error bound is developed for the ANNs approximators when the nuisance function belongs to a mixed Sobolev space. It is shown that the ANNs can alleviate the curse of dimensionality under this circumstance. The consistency and asymptotic normality of the proposed treatment effects estimators are further established, and a weighted bootstrap procedure for conducting inference is applied. The proposed methods are illustrated via simulation studies and a real data application.
