`tclust`

: Robust
Trimmed ClusteringThe package `tclust`

provides functions for robust trimmed
clustering. The methods are described in Garcia-Escudero (2008) doi:10.1214/07-AOS515,
Fritz et al. (2012) doi:10.18637/jss.v047.i12,
Garcia-Escudero et al. (2011) doi:10.1007/s11222-010-9194-z
and others.

The `tclust`

package is on CRAN (The Comprehensive R
Archive Network) and the latest release can be easily installed using
the command

```
install.packages("tclust")
library(tclust)
```

To install the latest stable development version from GitHub, you can pull this repository and install it using

```
## install.packages("remotes")
remotes::install_github("valentint/tclust", build_opts = c("--no-build-vignettes"))
```

Of course, if you have already installed `remotes`

, you
can skip the first line (I have commented it out).

Outlying data can heavily influence standard clustering methods. At
the same time, clustering principles can be useful when robustifying
statistical procedures. These two reasons motivate the development of
feasible robust model-based clustering approaches. Instead of trying to
“fit” noisy data, a proportion α of the most outlying observations is
trimmed. The `tclust`

package efficiently handles different
cluster scatter constraints. Graphical exploratory tools are also
provided to help the user make sensible choices for the trimming
proportion as well as the number of clusters to search for.

```
library(tclust)
#> Robust Trimmed Clustering (version 2.0-0)
data (M5data)
<- M5data[, 1:2]
x
<- tclust (x, k = 3, alpha = 0.1, restr.fact = 1,
clus.a restr = "eigen", equal.weights = TRUE)
<- tclust (x, k = 3, alpha = 0.1, restr.fact = 1,
clus.b equal.weights = TRUE)
<- tclust (x, k = 3, alpha = 0.1, restr.fact = 1,
clus.c restr = "deter", equal.weights = TRUE)
<- tclust (x, k = 3, alpha = 0.1, restr.fact = 50,
clus.d restr = "eigen", equal.weights = FALSE)
<- par (mfrow = c (2, 2))
pa plot (clus.a, main = "(a) tkmeans")
plot (clus.b, main = "(b) Gallegos and Ritter")
plot (clus.c, main = "(c) Gallegos")
plot (clus.d, main = "(d) tclust")
```

`par (pa)`

The trimmed k-means clustering method by Cuesta-Albertos, Gordaliza and Matran (1997) optimizes the k-means criterion under trimming a portion of the points:

```
library(tclust)
data (swissbank)
## Two clusters and 8% trimming level
<- tkmeans (swissbank, k = 2, alpha = 0.08)
clus
## Pairs plot of the clustering solution
pairs (swissbank, col = clus$cluster + 1)
```

```
# Two coordinates
plot (swissbank[, 4], swissbank[, 6], col = clus$cluster + 1,
xlab = "Distance of the inner frame to lower border",
ylab = "Length of the diagonal")
```

`plot (clus)`

If you experience any bugs or issues or if you have any suggestions
for additional features, please submit an issue via the *Issues*
tab of this repository. Please have a look at existing issues first to
see if your problem or feature request has already been discussed.

If you want to contribute to the package, you can fork this repository and create a pull request after implementing the desired functionality.

If you need help using the package, or if you are interested in collaborations related to this project, please get in touch with the package maintainer.