R/LearnerClustDBSCAN.R

In mlr3cluster: Cluster Extension for 'mlr3'

#' @title DBSCAN Clustering Learner
#'
#' @name mlr_learners_clust.dbscan
#' @include LearnerClust.R
#'
#' @description
#' DBSCAN (density-based spatial clustering of applications with noise) clustering.
#' Calls [dbscan::dbscan()] from package \CRANpkg{dbscan}.
#'
#' @templateVar id clust.dbscan
#' @template learner
#'
#' @references
#' `r format_bib("hahsler2019dbscan", "ester1996density")`
#'
#' @export
#' @template seealso_learner
#' @template simple_example
LearnerClustDBSCAN = R6Class(
  "LearnerClustDBSCAN",
  inherit = LearnerClust,
  public = list(
    #' @description
    #' Creates a new instance of this [R6][R6::R6Class] class.
    initialize = function() {
      param_set = ps(
        eps = p_dbl(0, tags = c("train", "required")),
        minPts = p_int(0L, default = 5L, tags = "train"),
        weights = p_uty(
          default = NULL,
          tags = "train",
          custom_check = crate(function(x) check_numeric(x, null.ok = TRUE))
        ),
        borderPoints = p_lgl(default = TRUE, tags = "train"),
        search = p_fct(c("kdtree", "linear", "dist"), default = "kdtree", tags = "train"),
        bucketSize = p_int(1L, default = 10L, tags = "train", depends = quote(search == "kdtree")),
        splitRule = p_fct(
          c("STD", "MIDPT", "FAIR", "SL_MIDPT", "SL_FAIR", "SUGGEST"),
          default = "SUGGEST",
          tags = "train",
          depends = quote(search == "kdtree")
        ),
        approx = p_dbl(default = 0, tags = "train")
      )

      super$initialize(
        id = "clust.dbscan",
        feature_types = c("logical", "integer", "numeric"),
        predict_types = "partition",
        param_set = param_set,
        properties = c("density", "exclusive", "partial"),
        packages = "dbscan",
        man = "mlr3cluster::mlr_learners_clust.dbscan",
        label = "DBSCAN"
      )
    }
  ),

  private = list(
    .train = function(task) {
      pv = self$param_set$get_values(tags = "train")
      data = as_numeric_matrix(task$data())
      m = invoke(dbscan::dbscan, x = data, .args = pv)
      m$data = data
      if (self$save_assignments) {
        self$assignments = m$cluster
      }
      m
    },

    .predict = function(task) {
      partition = invoke(
        predict,
        self$model,
        newdata = as_numeric_matrix(ordered_features(task, self)),
        data = self$model$data
      )
      list(partition = partition)
    }
  )
)

#' @include zzz.R
register_learner("clust.dbscan", LearnerClustDBSCAN)