Understanding Poisson-approximation resampling
Generalized linear models are an extension of the general linear model. Poisson
regression is a situation of generalized models. The dependent variable obeys
Poisson distribution.
Poisson sampling will be run on the Map of the MapReduce task because it occurs for
input data points. This doesn't guarantee that every data point will be considered into
the model, which is better than multinomial resampling of full datasets. But it will
guarantee the generation of independent samples by using N training input points.
different models for an ensemble classifier