Gaussian process regression with uncertainty sampling
Implementing Gaussian process regression using only numpy. Cholesky decomposition is used to stably compute the inverse matrix.
For more information, see http://www.gaussianprocess.org/gpml/chapters/RW.pdf
Uncertainty sampling (US) is one of the active learning methods to improve the accuracy of a model efficiently. Next observation point is the point with the maximum variance.
- noise variance: 1e-4
- kernel: RBF kernel
- length scale = 0.4
- variance = 1