readme.md

In dahcase/mbgstacking: Stacked generalization ensemble model for model-based geostatistics

Stacked Generalization for Model Based Geostatistics

Overview

Stacked generalization is a method of ensembling machine learning algorithms and/or statisical models such that the combination of a collection of children models performs better on predictive validity statistics than any individual component. In this implementation, a two stage approach is assumed: (1) a series of first stage (child) models are fit and (2) the child models are ensembled in a parent geostatistical model. It is also assumed that the user is a friend of IHME (although all are welcome) and has access to high performance computing. Specifically, the package is designed to take advantange of IHME's cluster computing infrastructure.

This work was inspired by Bhatt et al. (http://rsif.royalsocietypublishing.org/content/14/134/20170520.figures-only) and further details on the "why?" can be found there.

How To Install

devtools::install_github('dahcase/mbgstacking')

Getting Started

Tests

General Tips

dahcase/mbgstacking documentation built on May 20, 2019, 4:08 p.m.