Tree-Boosting for Spatial Data

 By Medium - 2021-03-20

Description

Combining gradient tree-boosting with Gaussian process models for modeling spatial data using GPBoost (Machine Learning; Spatial Statistics; Data Science; Artificial Intelligence; XGBoost; LightGBM)

Summary

  • GPBoost: The GPBoost algorithm is a boosting algorithm that iteratively learns the covariance parameters (aka hyperparameters) of the Gaussian process and adds a tree to the ensemble of trees using a gradient and/or a Newton boosting step which accounts for the spatial correlation.
  • In the GPBoost library, covariance parameters can be learned using (Nesterov accelerated) gradient descent or Fisher scoring (aka natural gradient descent), and trees are learned using the LightGBM library.
  • The GPBoost algorithm allows for relaxing this assumption in a flexible manner.

 

Topics

  1. Backend (0.29)
  2. Machine_Learning (0.28)
  3. NLP (0.13)

Similar Articles

Gaussian Process Regression From First Principles

By Medium - 2021-03-15

Gaussian Process Regression (GPR) is a remarkably powerful class of machine learning algorithms that, in contrast to many of today’s state-of-the-art machine learning models, relies on few parameters…

Facebook’s Prophet + Deep Learning = NeuralProphet

By Medium - 2020-12-10

While learning about time series forecasting, sooner or later you will encounter the vastly popular Prophet model, developed by Facebook. It gained lots of popularity due to the fact that it provides…