Description Usage Arguments Details Value Author(s) References See Also Examples
View source: R/changepoints_lin2_cl.R
This function detects change points for parameter changes of a linear AR(1) model with intercept based on simplicial depth. The model is defined by
Y_n = θ_1 Y_{n-1} + θ_0 + E_n
with med(E_n)=0. The main idea is based on estimators and confidence intervals based on simplicial depth as proposed by Kustosz (2016). Further, a preselection of candidates by a simplified depth notion is used. This is a version which uses a preset cluster which hast to be passed as argument when calling the function.
1 | changepoints_lin2_cl(y, level, bw, sw, plots, method, cluster, mincper, mincp)
|
y |
An observed AR(1) process with intercept. |
level |
The level for which the confidence intervals for the phase change detection are calculated. |
bw |
A bandwidth defining the sample size for sub-samples which are evaluated in the detection algorithm. |
sw |
A step-width to reduce the candidate points in which the change points are evaluated. |
plots |
A switch (TRUE and FALSE) to determine if plots shall be produced for each tested candidate |
method |
This parameter determines the method to detect change points. If |
cluster |
Initialized cluster for parallel computing. |
mincper |
Percentage to define relevant change-point clusters based on the length of the 'mincp' longest detected cluster. |
mincp |
The 'mincp' longest cluster is used to define a threshold for detected clusters which are considered for final break-point estimates. This implies, that (if enough clusters exist), the number also defines the minimal amount of detected breakpoints. |
For details see Kustosz (2016).
candidates |
A vector of observation indices in which the algorithm tested for change points. |
changepoints |
A vector indicating, if a change point was detected in the respective candidate. |
Kustosz, Christoph
Kustosz, C. (2016). Depth based estimators and tests for
autoregressive processes with application. Ph. D. thesis. TU Dortmund.
RandomARMod_lin2
, lin2_CI_cl
, convex_hull_plot
, convex_hull_intersect
, est_lin2
, dS_lin2
, dS_lin2_test
, dS1_lin2_test
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 | int1 <- 0.1
int2 <- 0.3
arp1 <- 1.015
arp2 <- 1.001
set.seed(123)
y1 <- RandomARMod_lin2(50, int1, arp1, 15, "0")
y2 <- RandomARMod_lin2(50, int2, arp2, y1[length(y1)], "0")
y <- c(y1,y2)
## Not run:
cl <- makeCluster(2)
CP1 <- changepoints_lin2_cl(y, 0.99, 25, 5, plots = FALSE, method = "m1", cl = cl)
CP2 <- changepoints_lin2_cl(y, 0.99, 25, 5, plots = FALSE, method = "m2", cl = cl)
stopCluster(cl)
par(mfrow = c(2, 1))
plot(y, type="l")
abline(v=CP1$candidates[CP1$changepoints==1], col=4,lty=4)
abline(v=CP2$candidates[CP2$changepoints==1], col=5,lty=5)
plot(y, type="l")
abline(v=CP1$rjumps, col=4,lty=4,lwd=2)
abline(v=CP2$rjumps, col=5,lty=5,lwd=3)
## End(Not run)
|
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.