http://arxiv.org/abs/1505.03036
We describe a method for removing the effect of confounders in order to reconstruct a latent quantity of interest. The method, referred to as half-sibling regression, is inspired by recent work in causal inference using additive noise models. We provide a theoretical justification and illustrate the potential of the method in a challenging astronomy application.
B. Scholkopf, D. Hogg, D. Wang, et. al.
Wed, 13 May 15
6/69
Comments: Extended version of a paper appearing in the Proceedings of the 32nd International Conference on Machine Learning, Lille, France, 2015