Documentation for Kriging?

[rkube]

Hi, The documentation for Kriging is confusing. What kind of model is used to calculate the covariance and what exactly do the parameters p and theta refer to in the model?

Im also unsure how the error is calculated. The tutorial calls the function std_error_at_point but this function is not documented in the package and which can't resolve the method:

julia> which(std_error_at_point, (typeof(kriging_surrogate),))
ERROR: no unique matching method found for the specified argument types
Stacktrace:
 [1] error(s::String)
   @ Base ./error.jl:35
 [2] _which
   @ ./reflection.jl:1498 [inlined]
 [3] _which
   @ ./reflection.jl:1496 [inlined]
 [4] which(tt::Any)
   @ Base ./reflection.jl:1523
 [5] which(f::Any, t::Any)
   @ Base ./reflection.jl:1514
 [6] top-level scope
   @ REPL[64]:1

[ChrisRackauckas]

What kind of model is used to calculate the covariance and what exactly do the parameters p and theta refer to in the model?

Kriging has a standard Gaussian interpretation for this. The standard stuff is all at https://en.wikipedia.org/wiki/Kriging

[Spinachboul]

@rkube I wanted to tell about the paramters p and theta.

• P: Refers to the power in the covariance function. It controls the smoothness of the resulting interpolated surface.
• Theta: Refers to the "length scale" which determines how quickly the correlation between the points decreases as the distance between them increases.

Adjusting p and theta allows the Kriging model to adapt to different spatial patterns and characteristics in the data. Choosing appropriate values for these parameters is crucial for obtaining accurate and reliable predictions or interpolations.

[Spinachboul]

Is this issue still open??