DRAFT: First basic implementation of GP

[mamarder]

Overview A first draft of a simple GP is implemented in ramsey/_src/gaussian_process/gaussian_process.py

An RBF Kernel is offered in ramsey/_src/gaussian_process/kernel.py In general the design is chosen so that a library consumer can implement own kernels.

An example on how to use the the module and some visualization is given in examples/gaussian_process/gaussian_process.py This examples generates plots as shown below.

The example generates data by using function from examples/gaussian_process/data.py

Limitations

• training the GP leads to numerical problems depending on the nature of the generative process, the number of training points and the initial parameters of the training
• only 1D inputs have been tested so far

[dirmeier]

Thanks! Looks great, I'll merge this now.