Nonparametric MOSUM wrapper (NP-MOJO) — npmojo

Wraps CptNonPar::np.mojo() (McGonigle and Cho, 2023): nonparametric moving-sum detection of changes in the marginal or joint distribution of a (possibly multivariate) time series, robust to serial dependence.

Usage

npmojo_wrapper(x, G = NULL, lag = 0, ...)

Arguments

x: A numeric vector or matrix (rows are time points).
G: Moving-window bandwidth. Defaults to max(20, 0.1 * n) observations.
lag: Time lag at which changes in the joint distribution are examined; 0 targets the marginal distribution. Defaults to 0.
...: Additional arguments passed to CptNonPar::np.mojo().

Value

A ggcpt object.

References

McGonigle ET, Cho H (2025). “Nonparametric data segmentation in multivariate time series via joint characteristic functions.” Biometrika, 112(2), asaf024.

Examples

res <- npmojo_wrapper(c(rnorm(100), rnorm(100, 3)))
res$changepoints
#> # A tibble: 2 × 2
#>      cp cp_value
#>   <int>    <dbl>
#> 1   100    0.369
#> 2   122    5.13