Fast implementation of the HDBSCAN (Hierarchical DBSCAN) and its related algorithms using Rcpp.
hdbscan(x, minPts, xdist = NULL,
gen_hdbscan_tree = FALSE,
gen_simplified_tree = FALSE)# S3 method for hdbscan
print(x, ...)
# S3 method for hdbscan
plot(x, scale="suggest",
gradient=c("yellow", "red"), show_flat = FALSE, ...)
a data matrix or a dist object.
integer; Minimum size of clusters. See details.
an optional precomputed dist object. May reduce computation time if supplied.
logical; should the robust single linkage tree be explicitly computed. (see cluster tree in Chaudhuri et al, 2010).
logical; should the simplified hierarchy be explicitly computed. (see Campello et al, 2013).
additional arguments are passed on to the appropriate S3 methods (such as plotting parameters).
integer; used to scale condensed tree based on the graphics device. Lower scale results in wider trees.
character vector; the colors to build the condensed tree coloring with.
logical; whether to draw boxes indicating the most stable clusters.
A object of class 'hdbscan' with the following components:
A integer vector with cluster assignments. Zero indicates noise points.
value of the minPts parameter.
The sum of the stability scores for each salient ('flat') cluster. Corresponds to cluster ids given the in 'cluster' member.
The 'probability' or individual stability of a point within its clusters. Between 0 and 1.
The outlier score (GLOSH) of each point.
An 'hclust' object of the HDBSCAN hierarchy.
Computes the hierarchical cluster tree representing density estimates along with the stability-based flat cluster extraction proposed by Campello et al. (2013). HDBSCAN essentially computes the hierarchy of all DBSCAN* clusterings, and then uses a stability-based extraction method to find optimal cuts in the hierarchy, thus producing a flat solution.
Additional, related algorithms including the "Global-Local Outlier Score from Hierarchies" (GLOSH) (see section 6 of Campello et al. 2015) outlier scores and ability to cluster based on instance-level constraints (see section 5.3 of Campello et al. 2015) are supported. The algorithms only need the parameter minPts
.
Note that minPts
not only acts as a minimum cluster size to detect, but also as a "smoothing" factor of the density estimates implicitly computed from HDBSCAN.
Martin Ester, Hans-Peter Kriegel, Joerg Sander, Xiaowei Xu (1996). A Density-Based Algorithm for Discovering Clusters in Large Spatial Databases with Noise. Institute for Computer Science, University of Munich. Proceedings of 2nd International Conference on Knowledge Discovery and Data Mining (KDD-96).
Campello, R. J. G. B.; Moulavi, D.; Sander, J. (2013). Density-Based Clustering Based on Hierarchical Density Estimates. Proceedings of the 17th Pacific-Asia Conference on Knowledge Discovery in Databases, PAKDD 2013, Lecture Notes in Computer Science 7819, p. 160.
Campello, Ricardo JGB, et al. "Hierarchical density estimates for data clustering, visualization, and outlier detection." ACM Transactions on Knowledge Discovery from Data (TKDD) 10.1 (2015): 5.
# NOT RUN {
## cluster the moons data set with HDBSCAN
data(moons)
res <- hdbscan(moons, minPts = 5)
res
plot(res)
plot(moons, col = res$cluster + 1L)
## DS3 from Chameleon
data("DS3")
res <- hdbscan(DS3, minPts = 50)
res
## Plot the simplified tree, highlight the most stable clusters
plot(res, show_flat = TRUE)
## Plot the actual clusters
plot(DS3, col=res$cluster+1L, cex = .5)
# }
Run the code above in your browser using DataLab