Robust Fusion Methods for Big Data
arXiv:1705.10157
Abstract
We address one of the important problems in Big Data, namely how to combine estimators from different subsamples by robust fusion procedures, when we are unable to deal with the whole sample.