Transformations

The aeon.transformations module contains classes for data transformations.

Transformations.

All (simple) transformers in aeon can be listed using the aeon.registry .all_estimators utility, using estimator_types=”transformer” tag.

Transformations are categorized as follows:

Category	Explanation	Example
Composition	Building blocks for pipelines, wrappers, adapters	Transformer pipeline
Series-to-tabular	Transforms series to tabular data	Length and mean
series-to-series	Transforms individual series to series	Differencing, detrending
Series-to-collection	transforms a series into a collection of time series	Bootstrap, sliding window
Collection	Transforms a collection of times series into a new collection of time series	Padding to equal length
Hierarchical	uses hierarchy information non-trivially	Reconciliation

Composition

Sklearn and pandas adapters

Tabularizer()

A transformer that turns time series collection into tabular data.

Series-to-tabular transformers

Series-to-tabular transformers transform individual time series to a vector of features, usually a vector of floats, but can also be categorical.

When applied to collections or hierarchical data, the transformation result is a table with as many rows as time series in the collection and a column for each feature.

Shapelets, wavelets and convolution

RandomShapeletTransform([...])	Random Shapelet Transform.
RandomDilatedShapeletTransform([...])	Random Dilated Shapelet Transform (RDST) as described in [R1a26faa97573-1], [R1a26faa97573-2].
SAST([lengths, stride, nb_inst_per_class, ...])	Scalable and Accurate Subsequence Transform (SAST).

Rocket([num_kernels, normalise, n_jobs, ...])	RandOm Convolutional KErnel Transform (ROCKET).
MiniRocket([num_kernels, ...])	MINImally RandOm Convolutional KErnel Transform (MiniRocket).
MiniRocketMultivariateVariable([...])	MINIROCKET (Multivariate, unequal length).
MultiRocket([num_kernels, ...])	Multi RandOm Convolutional KErnel Transform (MultiRocket).

DWTTransformer([n_levels])

Discrete Wavelet Transform Transformer.

Distance-based features

MatrixProfile([m])

Return the matrix profile and index profile for each time series of a dataset.

Dictionary-based features

PAA([n_segments])	Piecewise Aggregate Approximation Transformer (PAA).
SFA([word_length, alphabet_size, ...])	Symbolic Fourier Approximation (SFA) Transformer.
SAX([n_segments, alphabet_size, ...])	Symbolic Aggregate approXimation (SAX) transformer.

Signature-based features

SignatureTransformer([augmentation_list, ...])

Transformation class from the signature method.

Feature collections

These transformers extract larger collections of features.

TSFreshRelevantFeatureExtractor([...])	Transformer for extracting time series features via tsfresh.extract_features.
TSFreshFeatureExtractor([...])	Transformer for extracting time series features via tsfresh.extract_features.
Catch22([features, catch24, outlier_norm, ...])	Canonical Time-series Characteristics (Catch22).

Series transforms

These transformations apply a function element-wise.

Depending on the transformer, the transformation parameters can be fitted.

BoxCoxTransformer([bounds, method, sp])	Box-Cox power transform.
LogTransformer([offset, scale])	Natural logarithm transformation.

Detrending

Filtering and denoising

BKFilter([low, high, K])

Filter a times series using the Baxter-King filter.

DFTSeriesTransformer([r, sort])

Filter a times series using Discrete Fourier Approximation (DFT).

Slope

SlopeTransformer([n_intervals])

Piecewise slope transformation.

Segmentation

TSInterpolator(length)

Time series interpolator/re-sampler.

IntervalSegmenter([intervals])	Interval segmentation transformer.
RandomIntervalSegmenter([n_intervals, ...])	Random interval segmenter transformer.
SlidingWindowSegmenter([window_length])	Sliding window segmenter transformer.

Window-based series transforms

These transformers create a series based on a sequence of sliding windows.

HOG1DTransformer([n_intervals, n_bins, ...])

HOG1D transform.

ChannelScorer([classifier, proportion])	Channel scorer performs channel selection using a single channel classifier.
ElbowClassSum([distance, prototype_type, ...])	Elbow Class Sum (ECS) transformer to select a subset of channels/variables.
ElbowClassPairwise([distance, ...])	Elbow Class Pairwise (ECP) transformer to select a subset of channels.
RandomChannelSelector([p, random_state])	Selects a random proportion of channels.

Panel transformers

Panel transformers transform a panel of time series into a panel of time series.

A panel transformer is fitted on an entire panel, and not per series.

Equal length transforms

These transformations ensure all series in a panel have equal length

PaddingTransformer([pad_length, fill_value])

Pad unequal length time series to equal, fixed length.

TruncationTransformer([truncated_length])

Truncate unequal length time series to a lower bounds.

Dimension reduction

Series-to-Panel transformers

These transformers create a panel from a single series.

Outlier detection, changepoint detection

ClaSPTransformer([window_length, ...])

ClaSP (Classification Score Profile) Transformer.