Computation of log likelihood and AIC type information criteria for partitions given by breakpoints.
# S3 method for breakpointsfull
logLik(object, breaks = NULL, ...)
# S3 method for breakpointsfull
AIC(object, breaks = NULL, ..., k = 2)
# S3 method for breakpointsfull
LWZ(object, ...)
# S3 method for breakpoints
LWZ(object, ...)An object of class "logLik" or a simple vector containing
the AIC respectively.
an object of class "breakpoints" or "breakpointsfull".
if object is of class "breakpointsfull" the
number of breaks can be specified.
for LWZ, parameters passed to AIC.breakpointsfull. Unused in logLik and AIC.
the penalty parameter to be used, the default k = 2
is the classical AIC, k = log(n) gives the BIC, k = 0.299 * log(n)^2.1 gives the LWZ,
if n is the number of observations.
As for linear models the log likelihood is computed on a normal model and the degrees of freedom are the number of regression coefficients multiplied by the number of segments plus the number of estimated breakpoints plus 1 for the error variance.
If AIC or LWZ is applied to an object of class "breakpointsfull"
breaks can be a vector of integers and the AIC or LWZ for each corresponding
partition will be returned. By default the maximal number of breaks stored
in the object is used. See below for an example.
Liu, J., Wu, S., & Zidek, J. V. (1997). On segmented multivariate regression. Statistica Sinica, 497-525.
breakpoints
## Nile data with one breakpoint: the annual flows drop in 1898
## because the first Ashwan dam was built
data("Nile")
plot(Nile)
bp.nile <- breakpoints(Nile ~ 1)
summary(bp.nile)
plot(bp.nile)
## BIC of partitions with 0 to 5 breakpoints
plot(0:5, AIC(bp.nile, k = log(bp.nile$nobs)), type = "b")
## AIC
plot(0:5, AIC(bp.nile), type = "b")
## LWZ
plot(0:5, LWZ(bp.nile), type = "b")
## BIC, AIC, LWZ, log likelihood of a single partition
bp.nile1 <- breakpoints(bp.nile, breaks = 1)
AIC(bp.nile1, k = log(bp.nile1$nobs))
AIC(bp.nile1)
LWZ(bp.nile1)
logLik(bp.nile1)
Run the code above in your browser using DataLab