Gaussian processes (GPs) are very useful class of semi-parametric machine learning models. Before their use in more modern classification and regression tasks, they have been very successfully applied in searching for underground oil fields. GPs were called kriging models 1 back then, but the idea was the same. GPs belong to a general class of models known as kernel methods 2 . Kernel methods use something called the kernel function, denoted as

GSoC ends but the contribution continues This GSoC has been a great experience for me. Partly because of the technical aspects of the project, I loved coding the algorithms and contributing to an open-source project felt refreshing ! But most importantly because of my mentors Chris Fonnesbeck and Bill Engels. Their feedback was timely and precise, thank you ! I also found the PyMC community to be very helpful, considerate and welcoming !