Mixed Effects Machine Learning with GPBoost for Grouped and Areal Spatial Econometric Data
<p>The <a href="https://www.jmlr.org/papers/v23/20-322.html" rel="noopener ugc nofollow" target="_blank">GPBoost algorithm</a> extends linear mixed effects and Gaussian process models by replacing the linear fixed effects function with a non-parametric non-linear function modeled using tree-boosting. This article shows how the GPBoost algorithm implemented in the <code><a href="https://github.com/fabsig/GPBoost" rel="noopener ugc nofollow" target="_blank">GPBoost</a></code><a href="https://github.com/fabsig/GPBoost" rel="noopener ugc nofollow" target="_blank"> library</a> can be used for modeling data with a spatial and grouped structure. We demonstrate the functionality of the <code>GPBoost</code> library using European GDP data which is an example of areal spatial econometric data.</p>
<h2>Table of contents</h2>
<p>∘ Introduction<br />
· · Basic workflow of GPBoost<br />
· · Data description<br />
· ·Data loading and short visualization<br />
∘ Training a GPBoost model<br />
∘ Choosing tuning parameters<br />
∘ Model interpretation<br />
· ·Estimated random effects model<br />
· ·Spatial effect map<br />
· ·Understanding the fixed effects function<br />
∘ Extensions<br />
· ·Separate random effects for different time periods<br />
· ·Interaction between space and fixed effects predictor variables<br />
· ·Large data<br />
· ·Other spatial random effects models<br />
· ·(Generalized) linear mixed effects and Gaussian process models<br />
∘ References</p>
<h1>Introduction</h1>
<h2>Basic workflow of GPBoost</h2>
<p>Applying a GPBoost model (= combined tree-boosting and random effects / GP models) involves the following main steps:</p>
<p><a href="https://towardsdatascience.com/mixed-effects-machine-learning-with-gpboost-for-grouped-and-areal-spatial-econometric-data-b26f8bddd385">Read More</a></p>