Gaussian Process Regression Model (GPR) is one of the non-linear Regression Analysis .
We should not consider many parameters to build complicated non-linear model because GPR uses Kernel Method .
Output of GPR is not only the one line but also the distribution of the line.
"Process" in the name "Gaussian Process" is used for time siries data.
But mathmatical structure of Gaussian process is not limited for the time siries data. So there are methods using Gaussian process model for various type data.
"Gaussian Processes for Machine Learning": Carl Edward Rasmussen and Christopher K. I. Williams, MIT Press 2006
Not only regression but also classification.