Cross-validation summaries

Share:

Description

Generate a data frame of statistical values associated with cross-validation

Usage

1

Arguments

m.cv

data frame containing: the coordinates of data, prediction columns, prediction variance of cross-validation data points, observed values, residuals, zscore (residual divided by kriging standard error), and fold. If the rbf.tcv function is used, the prediction variance and zscore (residual divided by standard error) will have NA's

Value

data frame containing: mean prediction errors (MPE), average kriging standard error (ASEPE), root-mean-square prediction errors (RMSPE), mean standardized prediction errors (MSPE), root-mean-square standardized prediction errors (RMSSPE), mean absolute percentage prediction errors (MAPPE), coefficient of correlation of the prediction errors (CCPE), coefficient of determination (R2) and squared coefficient of correlation of the prediction errors (pseudoR2)

Examples

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
library(gstat)
data(meuse) 
coordinates(meuse) <- ~x+y 
m <- vgm(.59, "Sph", 874, .04) 

# leave-one-out cross validation: 
out <- krige.cv(log(zinc)~1, meuse, m, nmax = 40) 
criterio.cv(out)

# multiquadratic function
data(preci)
coordinates(preci)~x+y

# predefined eta
tab <- rbf.tcv(prec~x+y,preci,eta=1.488733, rho=0, n.neigh=9, func="M") 
criterio.cv(tab)

Want to suggest features or report bugs for rdrr.io? Use the GitHub issue tracker.