ROBINDEN: ROBINDEN (ROBust INitialization based on inverse DENsity...
In ktaucenters: Robust Clustering Procedures

Description Usage Arguments Details Value Note Author(s) References See Also Examples

View source: R/ROBINDEN.R

ROBINDEN searches for k initial cluster seeds for k-means-based clustering methods.

1	ROBINDEN(D, data, k, mp = 10)

`D`	A distance matrix calculated on `data`.
`data`	A data matrix with n observations and p variables.
`k`	The number of cluster centers to find.
`mp`	The number of the nearest neighbors to find dense regions by LOF, the default is 10.

The centers are the observations located in the most dense region and far away from each other at the same time. In order to find the observations in the highly dense region, ROBINPOINTDEN uses point density estimation (instead of Local Outlier Factor, Breunig et al (2000)), see more details.

`centers`	A numeric vector of `k` initial cluster centers corresponding to the k indices of observations.
`idpoints`	A real vector containing the inverse density values of each point (observation).

this is a slightly modified version of ROBIN algorithm implementation done by Sarka Brodinova <sarka.brodinova@tuwien.ac.at>.

Juan Domingo Gonzalez <juanrst@hotmail.com>

Hasan AM, et al. Robust partitional clustering by outlier and density insensitive seeding. Pattern Recognition Letters, 30(11), 994-1002, 2009.

lof

K=5;
nk=100
Z <- rnorm(2 * K * nk);
centers_aux <- -floor(K/2):floor(K/2)
mues <- rep(5*centers_aux,2*nk*K )
X <-  matrix(Z + mues, ncol=2)
# Generate sintetic outliers (contamination level 20%)
X[sample(1:(nk * K),(nk * K) * 0.2), ] <-matrix(runif((nk * K) * 0.2 * 2,
                                          3 * min(X), 3 * max(X)),
                                          ncol = 2, nrow = (nk * K) * 0.2)
res <- ROBINDEN(D =dist(X), data=X, k = K);
# plot the Initial centers found
plot(X)
points(X[res$centers,],pch=19,col=4,cex=2)