tsseg.algorithms.random package

Random — baseline detector for benchmarking.

Description

Emits uniformly random change points or state labels. Intended as a lower bound for benchmarks and for pipeline testing.

Two modes of operation:

• Unsupervised — draws a random n_change_points and/or n_states.

• Semi-supervised — given quantities, randomises locations/labels.

Type: state detection

Supervision: unsupervised or semi-supervised

Scope: univariate and multivariate

Parameters

Name	Type	Default	Description
semi_supervised	bool	False	Enable semi-supervised mode.
n_change_points	int / None	None	Number of change points to emit.
n_states	int / None	None	Number of distinct states.
random_state	int / None	None	Random seed for reproducibility.

Usage

from tsseg.algorithms import RandomDetector

detector = RandomDetector(n_change_points=5, random_state=42)
labels = detector.fit_predict(X)

Implementation: Origin: new code.

Reference: —

Submodules

tsseg.algorithms.random.detector module

Module contents