To understand the Gaussian Process
- as a prior over random functions,
- a posterior over functions given observed data,
- as a tool for spatial data modeling and computer experiments,
- and simply as a flexible "nonparametric" regression tool.
We'll see that, almost in spite of a technical (over) analysis of its properties, and sometimes strange vocabulary used to describe its features,
- it is a simple extension to the linear (regression) model.
- Knowing that is all it takes to make use of it as a nearly unbeatable regression tool when input–output relationships are relatively smooth.
- And even sometimes when they are not.