Selects sampling sites stratified by land cover class `lc`, scaling sample sizes proportionally while guaranteeing a minimum sample size per class. Optional normalization is applied prior to clustering.
Usage
site_selction(
df,
tot.sample.size,
min.sample.size = 50,
normalize = TRUE,
max.cluster.num = 6,
elbow.operation = TRUE
)Arguments
- df
Data frame or matrix containing feature columns, a land cover column `lc`, and a weight column `weights`, where each row is a sample and each column is a feature.
- tot.sample.size
Integer. Total target number of sample points to select across all land cover classes.
- min.sample.size
Integer. Minimum number of sample points allocated to each land cover class. (Default = 50).
- normalize
Logical. Whether to apply min-max scaling to numeric columns in `df`. (Default = TRUE).
- max.cluster.num
Integer. Maximum number of clusters to consider if calculating optimal cluster count. (Default = 6).
- elbow.operation
Logical. Whether to automatically calculate the proper cluster number based on the elbow location. (Default = TRUE). If FALSE, the `max.cluster.num` will be used directly as the cluster size.