Detects changepoints using Maximum Mean Discrepancy (MMD) in a Reproducing Kernel Hilbert Space. This nonparametric approach can detect changes in any aspect of the distribution.
kernel_cpd_detect(
data,
penalty = "bic",
kernel = "rbf",
bandwidth = 0,
min_segment = 5
)List with changepoints and kernel statistics
Numeric vector or matrix (rows = observations)
Penalty for changepoints (default: "bic")
Kernel function to use:
"rbf": Radial Basis Function (Gaussian) kernel
"linear": Linear kernel
"poly": Polynomial kernel
"laplacian": Laplacian kernel
Kernel bandwidth (0 = automatic median heuristic)
Minimum segment length
Arlot, S., Celisse, A., and Harchaoui, Z. (2019). A kernel multiple change-point algorithm via model selection. Journal of Machine Learning Research, 20(162), 1-56.