MADRID

class MADRID(min_length=8, max_length=None, step_size=1, train_test_split=None)[source]

Bases: BaseSeriesAnomalyDetector

MADRID multi-length discord anomaly detector.

MADRID is a discord discovery algorithm that finds time series anomalies of all lengths at once [1], rather than committing to a single subsequence length. Instead of committing to a single subsequence length, MADRID runs the left-discord matrix-profile method DAMP for every subsequence length in a candidate set and combines the length-normalised discord profiles into a single per-point anomaly score. It is the faster successor of the MERLIN discord discovery algorithm.

For each candidate length m the algorithm computes an approximate left matrix profile with DAMP (the distance from each subsequence to its nearest neighbour that starts strictly to its left), normalises it by sqrt(m) so that profiles of different lengths are comparable, and stores it as a row of a multi-length discord table. Because DAMP prunes aggressively, only the per-length best discords carry an exactness guarantee, so predict builds its pointwise scores from exactly those: each length’s top discord paints its length-normalised score onto the points that subsequence covers, combined across lengths with the maximum, and points covered by no top discord score zero. Every emitted value is one the algorithm actually evaluated. The complete multi-length output, including the approximate discord table, is available via predict_discords. Higher scores indicate more anomalous points.

Points before train_test_split form a warm-up (training) region that is only used as reference history and is always scored zero, mirroring DAMP’s left-discord definition.

Parameters:
min_lengthint, default=8

Minimum subsequence length in the candidate set. Must be at least 4.

max_lengthint or None, default=None

Maximum subsequence length in the candidate set. Must be at most half the length of the series. If None, the reference implementation’s automatic mode is used: the maximum length is derived from the warm-up region as split // 20 and about 50 candidate lengths are sampled from the range instead of sweeping every integer, following the algorithm’s parameter-free motivation.

step_sizeint, default=1

Step between consecutive candidate subsequence lengths when max_length is given: the candidate set is range(min_length, max_length + 1, step_size). Ignored in automatic mode.

train_test_splitint, float or None, default=None

Location of the split point between the warm-up (training) region and the region searched for anomalies. An int is used directly as the split index. A float in (0, 1) is interpreted as a fraction of the series length. If None, a warm-up of max(max_length, len(X) // 5) points is used. The split must satisfy max_length <= split <= len(X) - max_length.

Notes

Capabilities

Missing Values

No

Multithreading

No

Univariate

Yes

Multivariate

No

References

[1]

Yue Lu, Thirumalai Vinjamoor Akhil Srinivas, Takaaki Nakamura, Makoto Imamura and Eamonn Keogh, “Matrix Profile XXX: MADRID: A Hyper-Anytime and Parameter-Free Algorithm to Find Time Series Anomalies of All Lengths,” 2023 IEEE International Conference on Data Mining (ICDM), Shanghai, China, 2023, pp. 1199-1204.

Examples

>>> import numpy as np
>>> from aeon.anomaly_detection.series.distance_based import MADRID
>>> rng = np.random.default_rng(2)
>>> X = np.sin(np.linspace(0, 12 * np.pi, 120)) + rng.normal(0, 0.05, 120)
>>> X[70:78] = 2.5  # inject an anomalous flat segment
>>> detector = MADRID(min_length=6, max_length=12, train_test_split=24)
>>> scores = detector.fit_predict(X)
>>> bool(65 <= int(np.argmax(scores)) <= 78)
True

Methods

clone([random_state])

Obtain a clone of the object with the same hyperparameters.

fit(X[, y, axis])

Fit time series anomaly detector to X.

fit_predict(X[, y, axis])

Fit time series anomaly detector and find anomalies for X.

get_class_tag(tag_name[, raise_error, ...])

Get tag value from estimator class (only class tags).

get_class_tags()

Get class tags from estimator class and all its parent classes.

get_fitted_params([deep])

Get fitted parameters.

get_params([deep])

Get parameters for this estimator.

get_tag(tag_name[, raise_error, ...])

Get tag value from estimator class.

get_tags()

Get tags from estimator.

predict(X[, axis])

Find anomalies in X.

predict_discords(X)

Run MADRID and return the complete multi-length discord information.

reset([keep])

Reset the object to a clean post-init state.

set_fit_request(*[, axis])

Configure whether metadata should be requested to be passed to the fit method.

set_params(**params)

Set the parameters of this estimator.

set_predict_request(*[, axis])

Configure whether metadata should be requested to be passed to the predict method.

set_tags(**tag_dict)

Set dynamic tags to given values.
clone(random_state=None)[source]

Obtain a clone of the object with the same hyperparameters.

A clone is a different object without shared references, in post-init state. This function is equivalent to returning sklearn.clone of self. Equal in value to type(self)(**self.get_params(deep=False)).

Parameters:
random_stateint, RandomState instance, or None, default=None

Sets the random state of the clone. If None, the random state is not set. If int, random_state is the seed used by the random number generator. If RandomState instance, random_state is the random number generator.

Returns:
estimatorobject

Instance of type(self), clone of self (see above)

fit(X, y=None, axis=1)[source]

Fit time series anomaly detector to X.

If the tag fit_is_empty is true, this just sets the is_fitted tag to true. Otherwise, it checks self can handle X, formats X into the structure required by self then passes X (and possibly y) to _fit.

Parameters:
Xone of aeon.base._base_series.VALID_SERIES_INPUT_TYPES

The time series to fit the model to. A valid aeon time series data structure. See aeon.base._base_series.VALID_SERIES_INPUT_TYPES for aeon supported types.

yone of aeon.base._base_series.VALID_SERIES_INPUT_TYPES, default=None

The target values for the time series. A valid aeon time series data structure. See aeon.base._base_series.VALID_SERIES_INPUT_TYPES for aeon supported types.

axisint

The time point axis of the input series if it is 2D. If axis==0, it is assumed each column is a time series and each row is a time point. i.e. the shape of the data is (n_timepoints, n_channels). axis==1 indicates the time series are in rows, i.e. the shape of the data is (n_channels, n_timepoints).

Returns:
BaseSeriesAnomalyDetector

The fitted estimator, reference to self.

fit_predict(X, y=None, axis=1) → ndarray[source]

Fit time series anomaly detector and find anomalies for X.

Parameters:
Xone of aeon.base._base_series.VALID_SERIES_INPUT_TYPES

The time series to fit the model to. A valid aeon time series data structure. See aeon.base._base_series.VALID_INPUT_TYPES for aeon supported types.

yone of aeon.base._base_series.VALID_SERIES_INPUT_TYPES, default=None

The target values for the time series. A valid aeon time series data structure. See aeon.base._base_series.VALID_SERIES_INPUT_TYPES for aeon supported types.

axisint, default=1

The time point axis of the input series if it is 2D. If axis==0, it is assumed each column is a time series and each row is a time point. i.e. the shape of the data is (n_timepoints, n_channels). axis==1 indicates the time series are in rows, i.e. the shape of the data is (n_channels, n_timepoints).

Returns:
np.ndarray

A boolean, int or float array of length len(X), where each element indicates whether the corresponding subsequence is anomalous or its anomaly score.

classmethod get_class_tag(tag_name, raise_error=True, tag_value_default=None)[source]

Get tag value from estimator class (only class tags).

Parameters:
tag_namestr

Name of tag value.

raise_errorbool, default=True

Whether a ValueError is raised when the tag is not found.

tag_value_defaultany type, default=None

Default/fallback value if tag is not found and error is not raised.

Returns:
tag_value

Value of the tag_name tag in cls. If not found, returns an error if raise_error is True, otherwise it returns tag_value_default.

Raises:
ValueError

if raise_error is True and tag_name is not in self.get_tags().keys()

Examples

>>> from aeon.classification import DummyClassifier
>>> DummyClassifier.get_class_tag("capability:multivariate")
True
classmethod get_class_tags()[source]

Get class tags from estimator class and all its parent classes.

Returns:
collected_tagsdict

Dictionary of tag name and tag value pairs. Collected from _tags class attribute via nested inheritance. These are not overridden by dynamic tags set by set_tags or class __init__ calls.

get_fitted_params(deep=True)[source]

Get fitted parameters.

State required:

Requires state to be “fitted”.

Parameters:
deepbool, default=True

If True, will return the fitted parameters for this estimator and contained subobjects that are estimators.

Returns:
fitted_paramsdict

Fitted parameter names mapped to their values.

get_params(deep=True)

Get parameters for this estimator.

Parameters:
deepbool, default=True

If True, will return the parameters for this estimator and contained subobjects that are estimators.

Returns:
paramsdict

Parameter names mapped to their values.

get_tag(tag_name, raise_error=True, tag_value_default=None)[source]

Get tag value from estimator class.

Includes dynamic and overridden tags.

Parameters:
tag_namestr

Name of tag to be retrieved.

raise_errorbool, default=True

Whether a ValueError is raised when the tag is not found.

tag_value_defaultany type, default=None

Default/fallback value if tag is not found and error is not raised.

Returns:
tag_value

Value of the tag_name tag in self. If not found, returns an error if raise_error is True, otherwise it returns tag_value_default.

Raises:
ValueError

if raise_error is True and tag_name is not in self.get_tags().keys()

Examples

>>> from aeon.classification import DummyClassifier
>>> d = DummyClassifier()
>>> d.get_tag("capability:multivariate")
True
get_tags()[source]

Get tags from estimator.

Includes dynamic and overridden tags.

Returns:
collected_tagsdict

Dictionary of tag name and tag value pairs. Collected from _tags class attribute via nested inheritance and then any overridden and new tags from __init__ or set_tags.

predict(X, axis=1) → ndarray[source]

Find anomalies in X.

Parameters:
Xone of aeon.base._base_series.VALID_SERIES_INPUT_TYPES

The time series to fit the model to. A valid aeon time series data structure. See aeon.base._base_series.VALID_SERIES_INPUT_TYPES for aeon supported types.

axisint, default=1

The time point axis of the input series if it is 2D. If axis==0, it is assumed each column is a time series and each row is a time point. i.e. the shape of the data is (n_timepoints, n_channels). axis==1 indicates the time series are in rows, i.e. the shape of the data is (n_channels, n_timepoints).

Returns:
np.ndarray

A boolean, int or float array of length len(X), where each element indicates whether the corresponding subsequence is anomalous or its anomaly score.

predict_discords(X)[source]

Run MADRID and return the complete multi-length discord information.

Unlike predict, which reduces to one guaranteed score per time point, this exposes everything the algorithm computes, mirroring the reference implementation’s output.

Parameters:
Xnp.ndarray

One-dimensional time series.

Returns:
dict

lengths: the candidate subsequence lengths m; scores: the best (length-normalised) discord score per length; locations: each best discord’s start position; discord_table: the raw (n_lengths, n) table. Positions pruned by DAMP hold unrefined estimates, so treat the table as approximate everywhere except the returned discord locations; best_interval: (start, end) of the overall best discord, i.e. the location and extent of argmax(scores), matching the reference implementation’s best_discord_loc.

reset(keep=None)[source]

Reset the object to a clean post-init state.

After a self.reset() call, self is equal or similar in value to type(self)(**self.get_params(deep=False)), assuming no other attributes were kept using keep.

Detailed behaviour:
removes any object attributes, except:

hyper-parameters (arguments of __init__) object attributes containing double-underscores, i.e., the string “__”

runs __init__ with current values of hyperparameters (result of get_params)

Not affected by the reset are:

object attributes containing double-underscores class and object methods, class attributes any attributes specified in the keep argument temporary callback context attached by scikit-learn

Parameters:
keepNone, str, or list of str, default=None

If None, all attributes are removed except hyperparameters. If str, only the attribute with this name is kept. If list of str, only the attributes with these names are kept.

Returns:
selfobject

Reference to self.

Raises:
TypeError

If ‘keep’ is not a string or a list of strings.

set_fit_request(*, axis: bool | None | str = '$UNCHANGED$') → MADRID

Configure whether metadata should be requested to be passed to the fit method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

  • True: metadata is requested, and passed to fit if provided. The request is ignored if metadata is not provided.

  • False: metadata is not requested and the meta-estimator will not pass it to fit.

  • None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

  • str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:
axisstr, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED

Metadata routing for axis parameter in fit.

Returns:
selfobject

The updated object.

set_params(**params)

Set the parameters of this estimator.

The method works on simple estimators as well as on nested objects (such as Pipeline). The latter have parameters of the form <component>__<parameter> so that it’s possible to update each component of a nested object.

Parameters:
**paramsdict

Estimator parameters.

Returns:
selfestimator instance

Estimator instance.

set_predict_request(*, axis: bool | None | str = '$UNCHANGED$') → MADRID

Configure whether metadata should be requested to be passed to the predict method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

  • True: metadata is requested, and passed to predict if provided. The request is ignored if metadata is not provided.

  • False: metadata is not requested and the meta-estimator will not pass it to predict.

  • None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

  • str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:
axisstr, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED

Metadata routing for axis parameter in predict.

Returns:
selfobject

The updated object.

set_tags(**tag_dict)[source]

Set dynamic tags to given values.

Parameters:
**tag_dictdict

Dictionary of tag name and tag value pairs.

Returns:
selfobject

Reference to self.