Validating and comparing interpolation methods

library(potentiomap)
data("synthetic_wells")
p <- ps_make_points(synthetic_wells[1:16, ], "x", "y", "gw_elevation",
                    "well_id", "EPSG:26916")

Validation designs represent different prediction tasks. Spatial separation is appropriate when transfer to unsampled areas matters; random folds answer a different question.

v <- ps_validate(p, c("IDW", "TPS"), design = "kfold", folds = 3,
                 prediction_mode = "direct", seed = 12)

## [inverse distance weighted interpolation]
## [inverse distance weighted interpolation]
## Warning: 
## Grid searches over lambda (nugget and sill variances) with  minima at the endpoints: 
##   (GCV) Generalized Cross-Validation 
##    minimum at  right endpoint  lambda  =  0.0001771007 (eff. df= 9.500013 )
## [inverse distance weighted interpolation]

v$fold_manifest[, c("fold_id", "training_count", "validation_count")]

##                 fold_id training_count validation_count
## xmin  repeat_001_fold_2             11                5
## xmin1 repeat_001_fold_1             10                6
## xmin2 repeat_001_fold_3             11                5

comparison <- ps_compare_methods(v, metric = "rmse")
comparison$ranking[, c("method", "rmse", "finite_coverage", "rank")]

##   method      rmse finite_coverage rank
## 1    IDW 1.3198115               1    2
## 7    TPS 0.4293667               1    1

These scores are conditional on the recorded folds; they are not universal method rankings or automatic map accuracy.