scanP: Scan over different penalty 'P' values

In dpseg: Piecewise Linear Segmentation by Dynamic Programming

Description

Runs the dpseg recursion for different values of the penalty parameter P and returns a matrix with the used P values, the resulting number of segments and (optionally) the median of segment variance of residuals.

Usage

scanP(x, y, P, var = TRUE, use.matrix = TRUE, plot = TRUE,
  verb = 1, ...)

Arguments

x	x-values
y	y-values
P	vector of penalties P to scan
var	add the median of the variances of residuals of all segments to output (save time by var=FALSE)
use.matrix	use the stored scoring function matrix for more efficient scans; set this to FALSE if you run into memory problems
plot	plot results
verb	print progress messages
...	parameters for dpseg (except P)

Value

Returns a matrix with the penalties P in the first column, the number of segments in the second column and the median of variances in the third column.

Examples

x <- oddata$Time
y <- log(oddata$A5)
par(mai=c(par("mai")[1:3], par("mai")[2])) # to show right axis
sp <- scanP(x=x, y=y, P=seq(-.01,.1,length.out=50), plot=TRUE)

dpseg documentation built on Aug. 17, 2020, 5:06 p.m.