feat: added hdbscan and optics learners (#56)

* feat: added hdbscan and optics learners * docs: news update
mlr-org · Feb 29, 2024 · 0134af3 · 0134af3
1 parent 15ecf93
commit 0134af3
Show file tree

Hide file tree

Showing 36 changed files with 647 additions and 86 deletions.
diff --git a/.Rbuildignore b/.Rbuildignore
@@ -1,3 +1,4 @@
+^LICENSE$
 .ignore
 .editorconfig
 .gitignore

diff --git a/DESCRIPTION b/DESCRIPTION
@@ -58,12 +58,14 @@ Collate:
     'LearnerClustFanny.R'
     'LearnerClustFarthestFirst.R'
     'LearnerClustFeatureless.R'
+    'LearnerClustHDBSCAN.R'
     'LearnerClustHclust.R'
     'LearnerClustKKMeans.R'
     'LearnerClustKMeans.R'
     'LearnerClustMclust.R'
     'LearnerClustMeanShift.R'
     'LearnerClustMiniBatchKMeans.R'
+    'LearnerClustOPTICS.R'
     'LearnerClustPAM.R'
     'LearnerClustSimpleKMeans.R'
     'LearnerClustXMeans.R'

diff --git a/NAMESPACE b/NAMESPACE
@@ -24,12 +24,14 @@ export(LearnerClustEM)
 export(LearnerClustFanny)
 export(LearnerClustFarthestFirst)
 export(LearnerClustFeatureless)
+export(LearnerClustHDBSCAN)
 export(LearnerClustHclust)
 export(LearnerClustKKMeans)
 export(LearnerClustKMeans)
 export(LearnerClustMclust)
 export(LearnerClustMeanShift)
 export(LearnerClustMiniBatchKMeans)
+export(LearnerClustOPTICS)
 export(LearnerClustPAM)
 export(LearnerClustSimpleKMeans)
 export(LearnerClustXMeans)

diff --git a/NEWS.md b/NEWS.md
@@ -1,6 +1,8 @@
 # mlr3cluster (development version)
 
 * Add DBSCAN learner from 'fpc' package
+* Add HDBSCAN learner from 'dbscan' package
+* Add OPTICS learner from 'dbscan' package
 
 # mlr3cluster 0.1.8
 

diff --git a/R/LearnerClustAffinityPropagation.R b/R/LearnerClustAffinityPropagation.R
@@ -1,8 +1,6 @@
 #' @title Affinity Propagation Clustering Learner
 #'
 #' @name mlr_learners_clust.ap
-#' @include LearnerClust.R
-#' @include aaa.R
 #'
 #' @description
 #' A [LearnerClust] for Affinity Propagation clustering implemented in [apcluster::apcluster()].
@@ -26,7 +24,7 @@ LearnerClustAP = R6Class("LearnerClustAP",
     initialize = function() {
       param_set = ps(
         s = p_uty(tags = c("required", "train")),
-        p = p_uty(default = NA, tags = "train", custom_check = crate(function(x) check_numeric(x))),
+        p = p_uty(default = NA, tags = "train", custom_check = check_numeric),
         q = p_dbl(0, 1, tags = "train"),
         maxits = p_int(1L, default = 1000L, tags = "train"),
         convits = p_int(1L, default = 100L, tags = "train"),
@@ -79,4 +77,5 @@ LearnerClustAP = R6Class("LearnerClustAP",
   )
 )
 
+#' @include aaa.R
 learners[["clust.ap"]] = LearnerClustAP
diff --git a/R/LearnerClustAgnes.R b/R/LearnerClustAgnes.R
@@ -1,8 +1,6 @@
 #' @title Agglomerative Hierarchical Clustering Learner
 #'
 #' @name mlr_learners_clust.agnes
-#' @include LearnerClust.R
-#' @include aaa.R
 #'
 #' @description
 #' A [LearnerClust] for agglomerative hierarchical clustering implemented in [cluster::agnes()].
@@ -83,4 +81,5 @@ LearnerClustAgnes = R6Class("LearnerClustAgnes",
   )
 )
 
+#' @include aaa.R
 learners[["clust.agnes"]] = LearnerClustAgnes
diff --git a/R/LearnerClustCMeans.R b/R/LearnerClustCMeans.R
@@ -1,8 +1,6 @@
 #' @title Fuzzy C-Means Clustering Learner
 #'
 #' @name mlr_learners_clust.cmeans
-#' @include LearnerClust.R
-#' @include aaa.R
 #'
 #' @description
 #' A [LearnerClust] for fuzzy clustering implemented in [e1071::cmeans()].
@@ -24,7 +22,7 @@ LearnerClustCMeans = R6Class("LearnerClustCMeans",
     initialize = function() {
       param_set = ps(
         centers = p_uty(
-          tags = c("required", "train"), default = 2L, custom_check = crate(check_centers)
+          tags = c("required", "train"), default = 2L, custom_check = check_centers
         ),
         iter.max = p_int(1L, default = 100L, tags = "train"),
         verbose = p_lgl(default = FALSE, tags = "train"),
@@ -81,4 +79,5 @@ LearnerClustCMeans = R6Class("LearnerClustCMeans",
   )
 )
 
+#' @include aaa.R
 learners[["clust.cmeans"]] = LearnerClustCMeans
diff --git a/R/LearnerClustCobweb.R b/R/LearnerClustCobweb.R
@@ -1,8 +1,6 @@
 #' @title Cobweb Clustering Learner
 #'
 #' @name mlr_learners_clust.cobweb
-#' @include LearnerClust.R
-#' @include aaa.R
 #'
 #' @description
 #' A [LearnerClust] for Cobweb clustering implemented in [RWeka::Cobweb()].
@@ -57,4 +55,5 @@ LearnerClustCobweb = R6Class("LearnerClustCobweb",
   )
 )
 
+#' @include aaa.R
 learners[["clust.cobweb"]] = LearnerClustCobweb
diff --git a/R/LearnerClustDBSCAN.R b/R/LearnerClustDBSCAN.R
@@ -1,19 +1,20 @@
-#' @title Density-Based Clustering Learner
+#' @title Density-based Spatial Clustering of Applications with Noise (DBSCAN) Clustering Learner
 #'
 #' @name mlr_learners_clust.dbscan
-#' @include LearnerClust.R
-#' @include aaa.R
 #'
 #' @description
-#' A [LearnerClust] for density-based clustering implemented in [dbscan::dbscan()].
-#' The predict method uses [dbscan::predict.dbscan_fast()] to compute the
-#' cluster memberships for new data.
+#' DBSCAN (Density-based spatial clustering of applications with noise) clustering.
+#' Calls [dbscan::dbscan()] from \CRANpkg{dbscan}.
 #'
 #' @templateVar id clust.dbscan
 #' @template learner
-#' @template example
+#'
+#' @references
+#' `r format_bib("ester1996density")`
 #'
 #' @export
+#' @template seealso_learner
+#' @template example
 LearnerClustDBSCAN = R6Class("LearnerClustDBSCAN",
   inherit = LearnerClust,
   public = list(
@@ -24,7 +25,7 @@ LearnerClustDBSCAN = R6Class("LearnerClustDBSCAN",
         eps = p_dbl(0, tags = c("required", "train")),
         minPts = p_int(0L, default = 5L, tags = "train"),
         borderPoints = p_lgl(default = TRUE, tags = "train"),
-        weights = p_uty(tags = "train", custom_check = crate(function(x) check_numeric(x))),
+        weights = p_uty(tags = "train", custom_check = check_numeric),
         search = p_fct(levels = c("kdtree", "linear", "dist"), default = "kdtree", tags = "train"),
         bucketSize = p_int(1L, default = 10L, tags = "train"),
         splitRule = p_fct(
@@ -52,10 +53,7 @@ LearnerClustDBSCAN = R6Class("LearnerClustDBSCAN",
     .train = function(task) {
       pv = self$param_set$get_values(tags = "train")
       m = invoke(dbscan::dbscan, x = task$data(), .args = pv)
-      m = set_class(
-        list(cluster = m$cluster, eps = m$eps, minPts = m$minPts, data = task$data(), dist = m$dist),
-        c("dbscan_fast", "dbscan")
-      )
+      m = insert_named(m, list(data = task$data()))
       if (self$save_assignments) {
         self$assignments = m$cluster
       }
@@ -64,10 +62,11 @@ LearnerClustDBSCAN = R6Class("LearnerClustDBSCAN",
     },
 
     .predict = function(task) {
-      partition = predict(self$model, newdata = task$data(), self$model$data)
+      partition = invoke(predict, self$model, newdata = task$data(), data = self$model$data)
       PredictionClust$new(task = task, partition = partition)
     }
   )
 )
 
+#' @include aaa.R
 learners[["clust.dbscan"]] = LearnerClustDBSCAN
diff --git a/R/LearnerClustDBSCANfpc.R b/R/LearnerClustDBSCANfpc.R
@@ -1,17 +1,20 @@
-#' @title Density-Based Clustering Learner with fpc
+#' @title Density-based Spatial Clustering of Applications with Noise (DBSCAN) Clustering Learner
 #'
 #' @name mlr_learners_clust.dbscan_fpc
 #'
 #' @description
-#' A [LearnerClust] for density-based clustering implemented in [fpc::dbscan()].
-#' The predict method uses [fpc::predict.dbscan()] to compute the
-#' cluster memberships for new data.
+#' DBSCAN (Density-based spatial clustering of applications with noise) clustering.
+#' Calls [fpc::dbscan()] from \CRANpkg{fpc}.
 #'
 #' @templateVar id clust.dbscan_fpc
 #' @template learner
-#' @template example
+#'
+#' @references
+#' `r format_bib("ester1996density")`
 #'
 #' @export
+#' @template seealso_learner
+#' @template example
 LearnerClustDBSCANfpc = R6Class("LearnerClustDBSCANfpc",
   inherit = LearnerClust,
   public = list(
@@ -60,10 +63,7 @@ LearnerClustDBSCANfpc = R6Class("LearnerClustDBSCANfpc",
     .train = function(task) {
       pars = self$param_set$get_values(tags = "train")
       m = invoke(fpc::dbscan, data = task$data(), .args = pars)
-      m = set_class(
-        list(cluster = m$cluster, eps = m$eps, MinPts = m$MinPts, isseed = m$isseed, data = task$data()),
-        "dbscan"
-      )
+      m = insert_named(m, list(data = task$data()))
       if (self$save_assignments) {
         self$assignments = m$cluster
       }
@@ -72,10 +72,11 @@ LearnerClustDBSCANfpc = R6Class("LearnerClustDBSCANfpc",
     },
 
     .predict = function(task) {
-      partition = as.integer(predict(self$model, data = self$model$data, newdata = task$data()))
+      partition = as.integer(invoke(predict, self$model, data = self$model$data), newdata = task$data())
       PredictionClust$new(task = task, partition = partition)
     }
   )
 )
 
+#' @include aaa.R
 learners[["clust.dbscan_fpc"]] = LearnerClustDBSCANfpc
diff --git a/R/LearnerClustDiana.R b/R/LearnerClustDiana.R
@@ -1,8 +1,6 @@
 #' @title Divisive Hierarchical Clustering Learner
 #'
 #' @name mlr_learners_clust.diana
-#' @include LearnerClust.R
-#' @include aaa.R
 #'
 #' @description
 #' A [LearnerClust] for divisive hierarchical clustering implemented in [cluster::diana()].
@@ -64,4 +62,5 @@ LearnerClustDiana = R6Class("LearnerClustDiana",
   )
 )
 
+#' @include aaa.R
 learners[["clust.diana"]] = LearnerClustDiana
diff --git a/R/LearnerClustEM.R b/R/LearnerClustEM.R
@@ -1,8 +1,6 @@
 #' @title Expectation-Maximization Clustering Learner
 #'
 #' @name mlr_learners_clust.em
-#' @include LearnerClust.R
-#' @include aaa.R
 #'
 #' @description
 #' A [LearnerClust] for Expectation-Maximization clustering implemented in
@@ -68,4 +66,5 @@ LearnerClustEM = R6Class("LearnerClustEM",
   )
 )
 
+#' @include aaa.R
 learners[["clust.em"]] = LearnerClustEM
diff --git a/R/LearnerClustFanny.R b/R/LearnerClustFanny.R
@@ -1,8 +1,6 @@
 #' @title Fuzzy Analysis Clustering Learner
 #'
 #' @name mlr_learners_clust.fanny
-#' @include LearnerClust.R
-#' @include aaa.R
 #'
 #' @description
 #' A [LearnerClust] for fuzzy clustering implemented in [cluster::fanny()].
@@ -71,4 +69,5 @@ LearnerClustFanny = R6Class("LearnerClustFanny",
   )
 )
 
+#' @include aaa.R
 learners[["clust.fanny"]] = LearnerClustFanny
diff --git a/R/LearnerClustFarthestFirst.R b/R/LearnerClustFarthestFirst.R
@@ -1,8 +1,6 @@
 #' @title Farthest First Clustering Learner
 #'
 #' @name mlr_learners_clust.ff
-#' @include LearnerClust.R
-#' @include aaa.R
 #'
 #' @description
 #' A [LearnerClust] for Farthest First clustering implemented in [RWeka::FarthestFirst()].
@@ -67,4 +65,5 @@ LearnerClustFarthestFirst = R6Class("LearnerClustFF",
   )
 )
 
+#' @include aaa.R
 learners[["clust.ff"]] = LearnerClustFarthestFirst
diff --git a/R/LearnerClustFeatureless.R b/R/LearnerClustFeatureless.R
@@ -1,8 +1,6 @@
 #' @title Featureless Clustering Learner
 #'
 #' @name mlr_learners_clust.featureless
-#' @include LearnerClust.R
-#' @include aaa.R
 #'
 #' @description
 #' A simple [LearnerClust] which randomly (but evenly) assigns observations to
@@ -80,4 +78,5 @@ LearnerClustFeatureless = R6Class("LearnerClustFeatureless",
   )
 )
 
+#' @include aaa.R
 learners[["clust.featureless"]] = LearnerClustFeatureless
diff --git a/R/LearnerClustHDBSCAN.R b/R/LearnerClustHDBSCAN.R
@@ -0,0 +1,63 @@
+#' @title Hierarchical DBSCAN (HDBSCAN) Clustering Learner
+#'
+#' @name mlr_learners_clust.hdbscan
+#'
+#' @description
+#' HDBSCAN (Hierarchical DBSCAN) clustering.
+#' Calls [dbscan::hdbscan()] from \CRANpkg{dbscan}.
+#'
+#' @templateVar id clust.hdbscan
+#' @template learner
+#'
+#' @references
+#' `r format_bib("campello2013density")`
+#'
+#' @export
+#' @template seealso_learner
+#' @template example
+LearnerClustHDBSCAN = R6Class("LearnerClustHDBSCAN",
+  inherit = LearnerClust,
+  public = list(
+    #' @description
+    #' Creates a new instance of this [R6][R6::R6Class] class.
+    initialize = function() {
+      param_set = ps(
+        minPts = p_int(0L, tags = c("required", "train")),
+        gen_hdbscan_tree = p_lgl(default = FALSE, tags = "train"),
+        gen_simplified_tree = p_lgl(default = FALSE, tags = "train")
+      )
+
+      super$initialize(
+        id = "clust.hdbscan",
+        feature_types = c("logical", "integer", "numeric"),
+        predict_types = "partition",
+        param_set = param_set,
+        properties = c("partitional", "exclusive", "complete"),
+        packages = "dbscan",
+        man = "mlr3cluster::mlr_learners_clust.hdbscan",
+        label = "HDBSCAN Clustering"
+      )
+    }
+  ),
+  private = list(
+    .train = function(task) {
+      pv = self$param_set$get_values(tags = "train")
+      m = invoke(dbscan::hdbscan, x = task$data(), .args = pv)
+      m = insert_named(m, list(data = task$data()))
+
+      if (self$save_assignments) {
+        self$assignments = m$cluster
+      }
+
+      return(m)
+    },
+
+    .predict = function(task) {
+      partition = as.integer(invoke(predict, self$model, newdata = task$data(), data = self$model$data))
+      PredictionClust$new(task = task, partition = partition)
+    }
+  )
+)
+
+#' @include aaa.R
+learners[["clust.hdbscan"]] = LearnerClustHDBSCAN
diff --git a/R/LearnerClustHclust.R b/R/LearnerClustHclust.R
@@ -1,8 +1,6 @@
 #' @title Agglomerative Hierarchical Clustering Learner
 #'
 #' @name mlr_learners_clust.hclust
-#' @include LearnerClust.R
-#' @include aaa.R
 #'
 #' @description
 #' A [LearnerClust] for agglomerative hierarchical clustering implemented in [stats::hclust()].
@@ -82,4 +80,5 @@ LearnerClustHclust = R6Class("LearnerClustHclust",
   )
 )
 
+#' @include aaa.R
 learners[["clust.hclust"]] = LearnerClustHclust