ktaucenters_aux: ktaucenters_aux

In ktaucenters: Robust Clustering Procedures

Description

Robust Clustering algorithm based on centers, a robust and efficient version of KMeans.

Usage

ktaucenters_aux(X, K, centers, tolmin, NiterMax)

Arguments

X	A matrix of size n x p.
K	The number of clusters.
centers	matrix of size K x p containing the K initial centers, one at each matrix-row.
tolmin	tolerance parameter used for the algorithm stopping rule
NiterMax	a maximun number of iterations used for the algorithm stopping rule

Value

A list including the estimated K centers and labels for the observations

• centers: matrix of size K x p, with the estimated K centers.

• cluster: array of size n x 1 integers labels between 1 and K.

• tauPath: sequence of tau scale values at each iterations.

• Wni: numeric array of size n x 1 indicating the weights associated to each observation.

• emptyClusterFlag: a boolean value. True means that in some iteration there were clusters totally empty

• niter: number of iterations untill convergence is achived or maximun number of iteration is reached

• didistance of each observation to its assigned cluster-center

Note

Some times, if the initial centers are wrong, the algorithm converges to a non-optimal (local) solution. To avoid that, the algorithm must be run several times. This task is carried out by ktaucenters

References

Gonzalez, J. D., Yohai, V. J., & Zamar, R. H. (2019). Robust Clustering Using Tau-Scales. arXiv preprint arXiv:1906.08198.

See Also

ktaucenters

Examples

# Generate Sintetic data (three cluster well separated)
Z=rnorm(600);
mues=rep(c(0,10,20),200)
X= matrix(Z+mues,ncol=2)

# Applying the algortihm
sal = ktaucenters_aux(
X, K=3, centers=X[sample(1:300,3), ],
tolmin=1e-3, NiterMax=100)

#plot the results
plot(X,type="n")
points(X[sal$cluster==1,],col=1);
points(X[sal$cluster==2,],col=2);
points(X[sal$cluster==3,],col=3);

ktaucenters documentation built on Aug. 3, 2019, 9:03 a.m.