The DI of new rows of RAW predictor values, QC'd and scaled as the reference was.

Usage

aoa_di(aref, values)

Arguments

aref: From aoa_reference().
values: Matrix or data frame of raw values, columns in the model's order.

Value

The DI of each row; NA where a channel is missing after QC.

Examples

ex <- example_landscape()
store <- dsm_prepare(ex$profiles, target = "soc_stock", raster_dir = ex$raster_dir,
                     windows = c(3, 7), out_dir = file.path(tempdir(), "landscape"),
                     percentage = "^clay_pct$", transform = "log1p", n_cores = 1,
                     overwrite = TRUE, verbose = FALSE)
data <- dsm_load(store, windows = integer(0), verbose = FALSE)
predictors <- as.character(data$store$predictors)
qc <- utils::read.csv2(file.path(data$patch_dir, "qc_table.csv"))
qc <- qc[match(predictors, qc$predictor), ]
# a fitted model's scaling is its predictor_scaling.csv; here, the profiles' own
x <- as.matrix(data$points[, predictors])
scaling <- data.frame(predictor = predictors, center = colMeans(x),
                      scale = apply(x, 2, sd))
plan <- spatial_folds(data$store$meta, k = 3, block_size = 0.05)
aref <- aoa_reference(data$points, predictors, qc, scaling, plan)
aoa_di(aref, data$points[1:5, predictors])
#> [1] 0 0 0 0 0