These are the codes of the Graph Laplacian based Gaussian Process(GL-GP) algorithm and the datasets in the paper "GRAPH BASED GAUSSIAN PROCESSES ON RESTRICTED DOMAINS" by David B Dunson, Hau-Tieng Wu and Nan Wu
Explanation to files
1.CovMatrix.m
The Matlab code of the GL-GP for predictions.
- Nys.m
The code of Nystrom type extension. Suppose the GL-GP covariance parameters are determined by a small number of labeled points, the corresponding response variables, and a small number of unlabelled points. Nys.m can extend the prediction to a large number of unlabeled points without determining the GL-GP parameters again.
- hand.mat
The dataset in the Raynauld disease example.
- hand.txt
The explanation of the variables in the dataset "hand.mat".
- spiral example.txt
The Matlab code to generate a random dataset and a regression function on a spiral in R^2.
- spiral.mat
The dataset on the spiral that we use in the paper.
- two spheres example.txt
The Matlab code to generate a random dataset and a regression function on the spheres example in R^3.
- two spheres.mat
The dataset on the two spheres example that we use in the paper.
Acknowledgement
David B Dunson and Nan Wu are supported by the Lifeplan project. The Lifeplan project has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement No 856506).