Skip to contents

Environmental Similarity Between a Target Site and Candidate Sites

Source: R/alternative_sites.R
ss_alt_similarity.Rd

Computes a 0-1 similarity score between a target (e.g. an inaccessible sampling site) and a set of candidate sites, based on multivariate distance in environmental covariate space.

Usage

ss_alt_similarity(
  target_values,
  candidate_values,
  categorical_vars = NULL,
  weights = NULL,
  method = c("mahalanobis", "euclidean", "gower"),
  normalize = TRUE,
  missing_method = "pairwise"
)

Arguments

target_values

Environmental values at the target site: a named numeric vector, or a single-row data frame with the same columns as candidate_values.

candidate_values

Environmental values at candidate sites: a matrix or data frame, one row per candidate.

categorical_vars

Character vector of categorical column names. Only used by method = "gower", and excluded from the other methods.

weights

Optional named numeric vector of variable importance weights; unnamed weights must match length(target_values).

method

Distance metric: "mahalanobis" (default, accounts for covariance among variables), "euclidean", or "gower" (handles mixed continuous/categorical data).

normalize

Logical, whether to z-score normalize numeric variables before distance calculation. Default TRUE.

missing_method

How to handle missing values: "pairwise" (default, handled per-comparison), "listwise", "mean", or "median".

Value

Numeric vector of similarity scores in [0, 1], one per candidate row (1 = perfect match).

See also

ss_alt_candidates(), ss_alt_rank(), ss_alt_sites()

Examples

target <- c(dem = 4000, slope = 0.3, ndvi = 0.4)
candidates <- data.frame(
  dem = c(3990, 4200, 3500),
  slope = c(0.31, 0.5, 0.1),
  ndvi = c(0.41, 0.2, 0.6)
)
ss_alt_similarity(target, candidates)
#> Warning: Singular covariance matrix detected, added regularization
#> [1] 0.3679722 0.3678800 0.3678794