Compute length-scale such that averagely distant points have correlation p.

See Best (1999) "Bayesian models for spatially correlated disease and exposure data". When the average distance between points is sufficiently small, returns value one.

Usage

best_average(D, kernel = matern, p = 0.01, ...)

Arguments

D

Matrix of distances between points.

kernel

A kernel function, defaults to matern.

p

A percentage, defaults to 0.01.

...

Additional arguments to kernel.