RclusTool (version 0.91.61)

computeGap: Gap computation

Description

Estimate the number of clusters thanks to the gap computation.

Usage

computeGap(sim, Kmax)

Value

The function returns a list containing:

val

vector containing the eigenvalues of the similarity matrix.

gap

vector containing gap values between two successive eigenvalues.

Kmax

estimated number of clusters.

Arguments

sim

similarity matrix.

Kmax

maximal number of clusters.

Details

computeGap returns an estimated number of clusters

Examples

dat <- rbind(matrix(rnorm(100, mean = 0, sd = 0.3), ncol = 2), 
           matrix(rnorm(100, mean = 2, sd = 0.3), ncol = 2), 
           matrix(rnorm(100, mean = 4, sd = 0.3), ncol = 2))

sim <- computeGaussianSimilarity(dat, 1)
res <- computeGap(sim, Kmax = 20)

plot(res$val[1:20], type = "o", ann = FALSE, axes = FALSE)
abline(v = res$Kmax, col = "darkred")
abline(h = res$val[res$Kmax], col = "darkred")
axis(side = 1, at = c(seq(0,20,by=5), res$Kmax), 
     labels = c(seq(0,20,by=5), res$Kmax), cex.axis = .7)
axis(side = 2)
title("Automatic estimation of number of clusters - Gap method")
mtext("Number of clusters", side = 1, line = 3)
mtext("Eigenvalue", side = 2, line = 3)
box()