clustRviz aims to enable fast computation and easy visualization of
Convex Clustering solution paths.
You can install clustRviz from github with:
# install.packages("devtools")
devtools::install_github("DataSlingers/clustRviz")Note that RcppEigen (which clustRviz internally) triggers many
compiler warnings (which cannot be suppressed per CRAN
policies).
Many of these warnings can be locally suppressed by adding the line
CXX11FLAGS+=-Wno-ignored-attributes to your ~/.R/Makevars file.
clustRviz also depends on the development version of the gganimate
package, which is not yet on CRAN, but may be found
here.
There are two main entry points to the clustRviz package, the CARP
and CBASS functions, which perform convex clustering and convex
biclustering respectively. We demonstrate the use of these two functions
on a text minining data set, presidential_speech, which measures how
often the 44 U.S. presidents used certain words in their public
addresses.
library(clustRviz)
data(presidential_speech)
presidential_speech[1:6, 1:6]
#> amount appropri british cent commerci commission
#> Abraham Lincoln 3.433987 2.397895 1.791759 2.564949 2.708050 2.079442
#> Andrew Jackson 4.248495 4.663439 2.995732 1.945910 3.828641 3.218876
#> Andrew Johnson 4.025352 3.091042 2.833213 3.332205 2.772589 2.079442
#> Barack Obama 1.386294 0.000000 0.000000 1.386294 0.000000 0.000000
#> Benjamin Harrison 4.060443 4.174387 2.302585 4.304065 3.663562 3.465736
#> Calvin Coolidge 3.713572 4.094345 1.386294 3.555348 2.639057 1.609438We begin by clustering this data set, grouping the rows (presidents) into clusters:
carp_fit <- CARP(presidential_speech)
#> Pre-computing weights and edge sets
#> Computing Convex Clustering [CARP] Path
#> Post-processing
print(carp_fit)
#> CARP Fit Summary
#> ====================
#>
#> Algorithm: CARP (t = 1.05)
#> Fit Time: 0.148 secs
#> Total Time: 0.443 secs
#>
#> Number of Observations: 44
#> Number of Variables: 75
#>
#> Pre-processing options:
#> - Columnwise centering: TRUE
#> - Columnwise scaling: FALSE
#>
#> Weights:
#> - Source: Radial Basis Function Kernel Weights
#> - Distance Metric: Euclidean
#> - Scale parameter (phi): 0.01 [Data-Driven]
#> - Sparsified: 4 Nearest Neighbors [Data-Driven]The algorithmic regularization technique employed by CARP makes
computation of the whole solution path almost immediate.
We can examine the result of CARP graphically. We begin with a
standard dendrogram, with three clusters highlighted:
plot(carp_fit, type = "dendrogram", k = 3)
Examing the dendrogram, we see two clear clusters, consisting of pre-WWII and post-WWII presidents and Warren G. Harding as a possible outlier. Harding is generally considered one of the worst US presidents of all time, so this is perhaps not too surprising.
A more interesting visualization is the dynamic path visualization, whereby we can watch the clusters fuse as the regularization level is increased:
plot(carp_fit, type = "dynamic_path")
clustRviz also provides interactive dendrograms using the
heatmaply package:
plot(carp_fit, type = "js")
The use of CBASS for convex biclustering is similar, and we
demonstrate it here with a cluster heatmap, with the regularization set
to give 3 observation clusters:
cbass_fit <- CBASS(presidential_speech)
#> Pre-computing column weights and edge sets
#> Pre-computing row weights and edge sets
#> Computing Convex Bi-Clustering [CBASS] Path
#> Post-processing rows
#> Post-processing columns
plot(cbass_fit, k.row = 3)
Setting type = "js" gives the traditional cluster heatmap:
plot(cbass_fit, type = "js", k.row = 3)
By default, if a regularization level is specified, all plotting
functions in clustRviz will plot the clustered data. If the
regularization level is not specified, the raw data will be plotted
instead:
plot(cbass_fit, type = "js")
More details about the use and mathematical formulation of CARP and
CBASS may be found in the package documentation.