Compute DBSCAN clusters from a umapscan object

This function runs DBSCAN with the specified arguments to compute new clusters on a umapscan object. Only points which don't have an already validated cluster are taken into account.

compute_clusters(
  us,
  parent = "",
  noise_only = FALSE,
  eps,
  minPts,
  graph = TRUE,
  alpha = 1,
  ellipses = TRUE
)

Arguments

us	umapscan object
parent	name of the parent cluster
noise_only	only compute clusters for current 'Noise' points
eps	eps argument passed to dbscan::dbscan()
minPts	minPts argument passed to dbscan::dbscan()
graph	if TRUE, display a plot of the computed clusters
alpha	point transparency for clusters plot
ellipses	if TRUE, plot confidence ellipses around clusters

Value

Returns an updated umapscan object, and optionally displays a clusters plot.

See also

new_umapscan(), describe_clusters(), get_cluster_data(), rename_cluster(), plot_clusters(), get_clusters_membership(), remove_cluster()

Examples

library(dplyr)
#> 
#> Attachement du package : ‘dplyr’
#> The following objects are masked from ‘package:stats’:
#> 
#>     filter, lag
#> The following objects are masked from ‘package:base’:
#> 
#>     intersect, setdiff, setequal, union
iris_num <- iris %>% select_if(is.numeric)
us <- new_umapscan(iris_num, n_neighbors = 25, min_dist = 0.1, seed = 1337)
us <- compute_clusters(us, minPts = 3, eps = 0.5)
compute_clusters(us, minPts = 3, eps = 0.45, alpha = 1, parent = "3")
#> 
#> Call: new_umapscan(d = iris_num, n_neighbors = 25, min_dist = 0.1, 
#>     seed = 1337)
#> 
#> UMAP embeddings of a 150x4 data frame
#> 
#> Clusters :
#> 
#>     levelName
#> 1            
#> 2  ¦--1      
#> 3  ¦--2      
#> 4  ¦--3      
#> 5  ¦   °--3_1
#> 6  °--<Noise>