tsseg.algorithms.dynp package

Dynamic Programming (DynP) — exact optimal segmentation.

Description

Dynamic programming finds the exact minimum of the sum of segment costs by enumerating all admissible partitions. The user must specify the number of change points in advance (consider PELT or BinSeg when this is unknown).

Complexity is \(O(C\,K\,n^{2})\) where K is the number of change points, n the sample count and C the cost-function complexity. To reduce runtime, increase min_size and jump:

• min_size controls the minimum distance between change points.

• jump controls the grid of admissible positions (only multiples of jump).

Type: change point detection

Supervision: semi-supervised (n_cps required)

Scope: univariate and multivariate

Complexity: \(O(C\,K\,n^{2})\)

Parameters

Name	Type	Default	Description
n_cps	int	1	Number of change points to detect (>= 0).
model	str	"l2"	Ruptures cost model.
min_size	int	2	Minimum segment length.
jump	int	5	Grid step for candidate positions.
cost_params	dict / None	None	Extra arguments for the cost function.
axis	int	0	Time axis.

Usage

from tsseg.algorithms import DynpDetector

detector = DynpDetector(n_cps=3, model="l2")
labels = detector.fit_predict(X)

Implementation: Vendored from ruptures v1.1.8. BSD 2-Clause.

Reference: Auger & Lawrence (1989), Bulletin of Mathematical Biology.

Submodules

tsseg.algorithms.dynp.detector module

Module contents