Confidence intervals for jumpoints in stepmented regression

Computes confidence intervals for the changepoints (or jumpoints) in a fitted `stepmented' model.

Usage

# S3 method for class 'stepmented'
confint(object, parm, level=0.95, method=c("delta", "score", "gradient"), <!-- %var.diff=FALSE,  -->
    round=TRUE, cheb=FALSE, digits=max(4, getOption("digits") - 1),
    .coef=NULL, .vcov=NULL, ...)

Arguments

object

a fitted stepmented object.

parm

the stepmented variable of interest. If missing the first stepmented variable in object is considered.

level

the confidence level required, default to 0.95.

method

which confidence interval should be computed. One of "delta", "score", or "gradient". Can be abbreviated. Currently only "delta" allowed.

round

logical. Should the values (estimates and lower/upper limits) rounded to the smallest observed value?

cheb

logical. If TRUE, the confidence limits are computed using the Chebyshev inequality which yields conservative confidence intervals but it is 'robust' to the non-normality of the changepoint sampling distribution.

digits

controls the number of digits to print when returning the output.

.coef

The regression parameter estimates. If unspecified (i.e. NULL), it is computed internally by coef(object).

.vcov

The full covariance matrix of estimates. If unspecified (i.e. NULL), the covariance matrix is computed internally by vcov(object).

...

additional arguments passed to vcov.stepmented, namely k.

Details

confint.stepmented computes confidence limits for the changepoints. Currently the only option is 'delta', i.e. to compute the approximate covariance matrix via a smoothing approximation (see vcov.stepmented) and to build the limits using the standard Normal quantiles. Note that, the limits are rounded to the lowest observed value, thus the resulting confidence interval might not be symmetric if the stepmented covariate has not equispaced values.

Value

A matrix including point estimate and confidence limits of the breakpoint(s) for the stepmented variable possibly specified in parm.

Author

Vito M.R. Muggeo

Note

Currently only method='delta' is allowed.

See also

stepmented and lines.segmented to plot the estimated breakpoints with corresponding confidence intervals.

Examples

set.seed(10)
x<-1:100
z<-runif(100)
y<-2+2.5*(x>45)-1.5*(x>70)+z+rnorm(100)
o<-stepmented(y, npsi=2)

confint(o) #round=TRUE is default
#>            Est. CI(95%).low CI(95%).up
#> psi1.index   47          46         48
#> psi2.index   70          68         73
confint(o, round=FALSE)  
#>                Est. CI(95%).low CI(95%).up
#> psi1.index 47.62909    46.39268   48.86550
#> psi2.index 70.96822    68.46782   73.46862