We have improved the thinning algorithms by implementing two modifications of the kd-tree method: local kd-trees (local_kd_tree) and restricted search (k_estimation).
The round and hashing method (round_hash) now is deprecated and has been removed from the package because of presenting lower accuracy than the other distance-based methods.
The thinning method argument method now includes the three main thinning strategies implemented in GeoThinneR for better understanding and usability. These are: “distance”, “grid”, and “precision”.
The search_type argument has been included for distance-based thinning (method = "distance") to select between the search algorithms (“brute”, “kd_tree”, “local_kd_tree”, and “k_estimation”).
The long_col argument now is called lon_col for consistency with the lat_col argument.
The default column names (lon_col and `lat_col’) have been changed from “decimalLongitude” and “decimalLatitude” to “lon” and “lat”.
The n argument for the grid-based thinning (method = "grid") now lets you specify the number of points to retain per grid cell.
GeoThinneR 1.1.0
CRAN release: 2024-10-03
Removed non-CRAN dependency (rtree) and associated functions related to R-Trees structure thinning method (r_tree_thinning()).
