Wraps kcpRS::kcpRS() (Cabrieto et al., 2018; the KCP framework of
Arlot, Celisse and Harchaoui, 2019). The data are mapped to a running
statistic (mean, variance, autocorrelation, or correlation) computed on a
sliding window, and a Gaussian-kernel change point analysis with a
permutation significance test is run on the statistic. Detecting changes
in running correlations or variances captures higher-order changes that
mean-based methods miss. Multivariate input (matrix or data frame) is
supported.
Usage
kcp_wrapper(
x,
running_stat = c("mean", "var", "autocorr", "corr"),
wsize = 25,
nperm = 1000,
kmax = 10,
alpha = 0.05,
seed = NULL,
...
)Arguments
- x
A numeric vector, matrix, or data frame (columns are variables).
- running_stat
Which running statistic to monitor:
"mean","var","autocorr", or"corr"(correlation requires at least two columns). Defaults to"mean".- wsize
Sliding window size for the running statistic. Defaults to
25.- nperm
Number of permutations for the significance test. Defaults to
1000.- kmax
Maximum number of changepoints considered. Defaults to
10.- alpha
Significance level of the permutation test. Defaults to
0.05.- seed
Optional seed for reproducibility of the permutation test.
- ...
Additional arguments passed to
kcpRS::kcpRS().
Value
A ggcpt object. Reported locations refer to the centre of
the sliding window in which the change occurs. The series must be at
least wsize long to form one window. Constant coordinates make
every running statistic NA, so they are dropped (with a warning)
before detection and an all-constant input returns an empty result.
References
Arlot S, Celisse A, Harchaoui Z (2019). “A kernel multiple change-point algorithm via model selection.” Journal of Machine Learning Research, 20(162), 1–56.
Cabrieto J, Adolf J, Tuerlinckx F, Kuppens P, Ceulemans E (2018). “Detecting long-lived autodependency changes in a multivariate system via change point detection and regime switching models.” Scientific Reports, 8, 15637.