Concept克里金
Kriging
A surrogate model method based on Gaussian process regression that assumes noise-free observations, so that the interpolation surface passes strictly through all sample points. It is commonly used in spatial statistics and optimization.
Timeline (last 90 days)
Oct 14
Stochastic kriging (SK) models each observation as the true function value plus random noise, and the predictor outputs a mean estimate and a posterior variance; this differs from ordinary kriging, which assumes noise-free observations and strictly interpolates the sample points.
Unverified50%