Welcome to skchange

A python library for fast change point and segment anomaly detection. The library is designed to be compatible with sktime. Numba is used for computational speed.

Installation

The library can be installed via pip:

pip install skchange

Requires python versions >= 3.10, < 3.14.

For better computational performance, it is recommended to install skchange with numba:

pip install skchange[numba]

Key features

• Fast: Numba is used for performance.

• Easy to use: Follows the conventions of sktime and scikit-learn.

• Easy to extend: Make your own detectors by inheriting from the base class templates. Create custom detection scores and cost functions.

• Segment anomaly detection: Detect intervals of anomalous behaviour in time series data.

• Subset anomaly detection: Detect intervals of anomalous behaviour in time series data, and infer the subset of variables that are responsible for the anomaly.

Mission

The goal of skchange is to provide a library for fast and easy-to-use changepoint-based algorithms for change and anomaly detection. The primary focus is on modern methods in the statistical literature.

Example

from skchange.anomaly_detectors import CAPA
from skchange.anomaly_scores import L2Saving
from skchange.compose.penalised_score import PenalisedScore
from skchange.datasets import generate_piecewise_normal_data
from skchange.penalties import make_linear_chi2_penalty

df = generate_piecewise_normal_data(
    means=[0, 8, 0, 5],
    lengths=[100, 20, 130, 50],
    proportion_affected=[1.0, 0.1, 1.0, 0.5],
    n_variables=10,
    seed=1,
)

score = L2Saving()  # Looks for segments with non-zero means.
penalty = make_linear_chi2_penalty(score.get_model_size(1), df.shape[0], df.shape[1])
penalised_score = PenalisedScore(score, penalty)
detector = CAPA(penalised_score, find_affected_components=True)
detector.fit_predict(df)

        ilocs  labels         icolumns
0  [100, 120)       1              [0]
1  [250, 300)       2  [2, 0, 3, 1, 4]

Licence

This project is a free and open-source software licensed under the BSD 3-clause license.