divide: Divide a stream of events into distinct continguous clusters

In zatonovo/kingsmen: A technique for detecting regime change in irregular time series

Description

This method is less susceptible to noise than a direct clustering approach using only interarrival rates.

Arguments

event	An Event object
method	Passed to hclust to specify the clustering algorithm
levels	Number of levels to use when cutting the tree
plot	Specify whether a plot be generated
...	Additional arguments to pass to partition

Value

The result of divide is an EventGroup object. This can be summarized using the standard summary function.

Usage

divide EventGroup divide(event, method='complete', levels=1, plot=TRUE, ...)

Details

Irregular time series are challenging to work with since many statistical approaches require regular observation intervals.

Author(s)

Brian Lee Yung Rowe

Examples

d <- Sys.Date() + cumsum(round(c(rnorm(20,15,6), rnorm(20,25,10))))
e <- Event(d, abs(rnorm(length(d))))
g <- divide(e, plot=FALSE)
summary(g)

zatonovo/kingsmen documentation built on May 4, 2019, 9:11 p.m.