order_counts: Order soil class counts

In obrl-soil/dsmartr: dsmartr: An R implementation of the DSMART algorithm

Order soil class counts

Description

Rearranges soil class predictions into descending order, handling ties sensibly.

Usage

order_counts(input = NULL, n_classes = NULL)

Arguments

input	A vector of integers n model runs long, where the integers refer to counts of soil classes.
n_classes	Integer; the total number of soil classes available to predict.

Value

An atomic vector of probabilities, or a vector of NA values of the same length. Ties are shuffled at random, e.g. for 20 model runs, e.g. 0 5 8 0 becomes either 3 2 1 4 or 3 2 4 1. These values should correspond to a lookup table ID column, so they can be linked to a soil class name.

Note

This is a helper function and not widely applicable. It is expected to be used in concert with raster::calc(), applied to a stack of tabulated model realisations produced by dsmartr::collate().

Examples

## Not run: 
# 100 model runs, from a cell in the middle of the heronvale demo dataset
cell_19291 <- c(29, 7, 7, 7, 12, 9, 29, 29, 29, 3, 16, 15, 9, 21, 24, 2, 9,
15, 34, 9, 24, 34, 24, 34, 21, 24, 15, 15, 9, 21, 21, 7, 7, 34, 34, 15, 15,
21, 7, 30, 32, 21, 29, 21, 24, 29, 31, 12, 24, 12, 21, 15, 15, 29, 12, 7, 21,
12, 7, 34, 9, 32, 7, 7, 32, 32, 12, 34, 29, 7, 3, 15, 15, 29, 7, 3, 34, 9, 9,
33, 32, 21, 34, 29, 33, 7, 7, 31, 7, 7, 24, 21, 15, 7, 24, 29, 21, 31, 29,
15, 7)
preds_19291 <- count_predictions(cell_19291)
ordered_preds <- order_counts(preds_19291)

## End(Not run)

obrl-soil/dsmartr documentation built on Feb. 1, 2024, 10:57 p.m.