MSclust.p.est: Estimate stabilising parameter p for mean-shift clustering of...
In ClairBee/AS.circular: ArchStats: circular functions

Description Usage Arguments Value Examples

Automation of graphical method of estimating stabilising parameter to be used in mean-shift clustering algorithms. Calculates correlation of kernel density estimate for successive values of p, selecting as the optimum the maximum value of p that is less than 1 when rounded to a certain degree of precision.

1	MSclust.p.est(data, dp = 4, plot = T)

`data`	Set of angular data to be clustered.
`dp`	How many decimal places of convergence is required before p is accepted?
`plot`	Boolean: plot the correlation curve or not? Default is T.

Suggested value of p to be used in mean-shift clustering functions such as MSclust.NB

1
2
3

ex1 <- c(rvonmises(120 * 0.3, mu = circular(pi/2), kappa = 10),
         rvonmises(120 * 0.7, mu = circular(pi), kappa = 3))
p <- MSclust.p.est(ex1, dp = 4)