double.hinge {chngpt}

Fit Double Hinge Models

Description

Fit double hinge models.

Usage

double.hinge(x, y, lower.y = NULL, upper.y = NULL,
         var.type = c("none", "bootstrap"), ci.bootstrap.size =
         1000, alpha = 0.05, save.boot = TRUE, ncpus = 1, 
         boot.ci.type=c("percentile","symmetric"))

## S3 method for class 'double.hinge'
 plot(x, which = NULL, xlim = NULL, 
 lwd = 2, lcol = "red",
 lty = 1, add.points = TRUE, add.ci = TRUE, breaks =
 20, mark.chngpt = FALSE, xlab = NULL, ylab = NULL,
 ...) 
## S3 method for class 'double.hinge'
 fitted(object, ...) 
## S3 method for class 'double.hinge'
 residuals(object, ...) 

Arguments

object

x

x

x

y

y

lower.y

lower.y

upper.y

upper.y

var.type

var.type

boot.ci.type

var.type

ci.bootstrap.size

ci.bootstrap.size

alpha

alpha

save.boot

save.boot

ncpus

ncpus

lcol

ncpus

lwd

ncpus

which

x

xlim

x

lty

x

add.points

x

add.ci

x

breaks

x

mark.chngpt

x

xlab

x

ylab

x

...

arguments passed along

Details

If lower.y and upper.y are not supplied, min(y) is taken as the function value when x is less than or equal to the first threshold, and max(y) is taken as the function value when x is greater than or equal to the second threshold.

If the function is expected to be decreasing between the two thresholds, lower.y and upper.y should be supplied to ensure the correct fit.

mse is residual sum of squares