Welcome to our website. The aim of this project is to introduce Gaussian Process (GP) to students, research scholars and others who are interested in this state-of-the-art Non-parametric Bayesian method. To this end, we first introduce the basic notion of Gaussian Process (GPs) here, and then, provide for coding examples to expose the students to develop their own GPs. We begin by developing simple 1D GPs using self-generated dataset to analyse the GPs in 1 dimension first. Later, these will be extended to 2D GPs.

By the end of this short crash course, we hope that you will:

• Have an understanding of using GPs as prior for generating posteriors for given set of observations.
• Use GPs for spatial data modeling, preferably on real environmental datasets that are available publicly.

Things you need to go through this project:

• Thirst for Machine Learning Models.
• Basic understanding of Bayesian Inference.
• Prior exposure to Linear Regression Methods.
• Basic knowledge of Python 2.7 to try the code provided.

References:

• Rasmussen, Carl Edward, and Christopher KI Williams. "Gaussian processes for machine learning. 2006." The MIT Press, Cambridge, MA, USA 38 (2006): 715-719.
• http://bobby.gramacy.com/r_packages/
• https://www.cs.toronto.edu/~frossard/post/linear_regression/
• http://krasserm.github.io/2018/03/19/gaussian-processes/
• Snelson, Edward, and Zoubin Ghahramani. "Sparse Gaussian processes using pseudo-inputs." Advances in neural information processing systems. 2006.