DetectorPipeline

Pipeline for time series anomaly, changepoint detection, segmentation.

Quickstart

from sktime.detection.compose import DetectorPipeline

estimator = DetectorPipeline(steps)

Parameters(1)

stepslist of sktime transformers and detectors, or

list of tuples (str, estimator) of sktime transformers or detectors. The list must contain exactly one forecaster. These are “blueprint” transformers resp forecasters, detector/transformer states do not change when fit is called.

Examples

>>> import numpy as np
>>> import pandas as pd
>>> from sktime.detection.lof import SubLOF
>>> from sktime.transformations.detrend import Detrender
>>> 
>>> n = 100
>>> x = pd. Series (np. linspace (0, 5, n) + np. random. normal (0, 0.1, size = n))
>>> x. at [50 ] = 100
>>> 
>>> pipeline = Detrender () * SubLOF (n_neighbors = 5, window_size = 5, novelty = True)
>>> pipeline. fit (x) DetectorPipeline(
... )
>>> y_hat = pipeline. transform (x)