Time series classification#

The sktime.classification module contains algorithms and composition tools for time series classification.

All classifiers in sktime can be listed using the sktime.registry.all_estimators utility, using estimator_types="classifier", optionally filtered by tags. Valid tags can be listed using sktime.registry.all_tags.

Composition#

ClassifierPipeline(classifier, transformers)

Pipeline of transformers and a classifier.

ColumnEnsembleClassifier(estimators[, ...])

Applies estimators to columns of an array or pandas DataFrame.

SklearnClassifierPipeline(classifier, ...)

Pipeline of transformers and a classifier.

Ensembles#

BaggingClassifier(estimator[, n_estimators, ...])

Bagging ensemble of time series classifiers.

ComposableTimeSeriesForestClassifier([...])

Time Series Forest Classifier as described in [R880353f3a8bd-1].

WeightedEnsembleClassifier(classifiers[, ...])

Weighted ensemble of classifiers with fittable ensemble weight.

Deep learning#

CNNClassifier([n_epochs, batch_size, ...])

Time Convolutional Neural Network (CNN), as described in [Rd06fa525366f-1].

FCNClassifier([n_epochs, batch_size, ...])

Fully Connected Neural Network (FCN), as described in [R4ddccc18e565-1].

LSTMFCNClassifier([n_epochs, batch_size, ...])

Implementation of LSTMFCNClassifier from Karim et al (2019) [1].

InceptionTimeClassifier([n_epochs, ...])

InceptionTime Deep Learning Classifier.

MACNNClassifier([n_epochs, batch_size, ...])

Multi-Scale Attention Convolutional Neural Classifier, as described in [R18bfe0ce893e-1].

MLPClassifier([n_epochs, batch_size, ...])

Multi Layer Perceptron Network (MLP), as described in [R2c17e28b4d16-1].

MCDCNNClassifier([n_epochs, batch_size, ...])

Multi Channel Deep Convolutional Neural Classifier, as described in [Re4d070570206-1].

ResNetClassifier([n_epochs, callbacks, ...])

Residual Neural Network as described in [1].

SimpleRNNClassifier([num_epochs, ...])

Simple recurrent neural network.

TapNetClassifier([n_epochs, batch_size, ...])

Time series attentional prototype network (TapNet), as described in [Rdbe61d137f31-1].

Dictionary-based#

BOSSEnsemble([threshold, max_ensemble_size, ...])

Ensemble of Bag of Symbolic Fourier Approximation Symbols (BOSS).

ContractableBOSS([n_parameter_samples, ...])

Contractable Bag of Symbolic Fourier Approximation Symbols (cBOSS).

IndividualBOSS([window_size, word_length, ...])

Single bag of Symbolic Fourier Approximation Symbols (IndividualBOSS).

IndividualTDE([window_size, word_length, ...])

Single TDE classifier, an extension of the Bag of SFA Symbols (BOSS) model.

MUSE([anova, variance, bigrams, window_inc, ...])

MUSE (MUltivariate Symbolic Extension).

TemporalDictionaryEnsemble([...])

Temporal Dictionary Ensemble (TDE).

WEASEL([anova, bigrams, binning_strategy, ...])

Word Extraction for Time Series Classification (WEASEL).

Distance-based#

ElasticEnsemble([distance_measures, ...])

The Elastic Ensemble (EE).

KNeighborsTimeSeriesClassifier([...])

KNN Time Series Classifier.

ProximityForest([random_state, ...])

Proximity Forest classifier.

ProximityStump([random_state, ...])

Proximity Stump class.

ProximityTree([random_state, get_exemplars, ...])

Proximity Tree class.

ShapeDTW([n_neighbors, subsequence_length, ...])

ShapeDTW classifier.

Dummy#

DummyClassifier([strategy, random_state, ...])

DummyClassifier makes predictions that ignore the input features.

Early classification#

ProbabilityThresholdEarlyClassifier([...])

Probability Threshold Early Classifier.

TEASER([estimator, one_class_classifier, ...])

Two-tier Early and Accurate Series Classifier (TEASER).

Feature-based#

Catch22Classifier([outlier_norm, ...])

Canonical Time-series Characteristics (catch22) classifier.

FreshPRINCE([default_fc_parameters, ...])

Fresh Pipeline with RotatIoN forest Classifier.

MatrixProfileClassifier([...])

Martrix Profile (MP) classifier.

RandomIntervalClassifier([n_intervals, ...])

Random Interval Classifier.

SignatureClassifier([estimator, ...])

Classification module using signature-based features.

SummaryClassifier([summary_functions, ...])

Summary statistic classifier.

TSFreshClassifier([default_fc_parameters, ...])

Time Series Feature Extraction based on Scalable Hypothesis Tests classifier.

Hybrid#

HIVECOTEV1([stc_params, tsf_params, ...])

Hierarchical Vote Collective of Transformation-based Ensembles (HIVE-COTE) V1.

HIVECOTEV2([stc_params, drcif_params, ...])

Hierarchical Vote Collective of Transformation-based Ensembles (HIVE-COTE) V2.

Interval-based#

CanonicalIntervalForest([n_estimators, ...])

Canonical Interval Forest Classifier (CIF).

DrCIF([n_estimators, n_intervals, ...])

Diverse Representation Canonical Interval Forest Classifier (DrCIF).

RandomIntervalSpectralEnsemble([...])

Random Interval Spectral Ensemble (RISE).

SupervisedTimeSeriesForest([n_estimators, ...])

Supervised Time Series Forest (STSF).

TimeSeriesForestClassifier([min_interval, ...])

Time series forest classifier.

Kernel-based#

TimeSeriesSVC([kernel, kernel_params, ...])

Support Vector Classifier, for time series kernels.

Arsenal([num_kernels, n_estimators, ...])

Arsenal ensemble.

RocketClassifier([num_kernels, ...])

Classifier wrapped for the Rocket transformer using RidgeClassifierCV.

Shapelet-based#

ShapeletTransformClassifier([...])

A shapelet transform classifier (STC).

MrSEQL([seql_mode, symrep, custom_config])

MrSEQL = Multiple Representations Sequence Learning classification model.

MrSQM([strat, features_per_rep, ...])

MrSQM = Multiple Representations Sequence Miner.

sklearn#

ContinuousIntervalTree([max_depth, ...])

Continuous interval tree (CIT) vector classifier (aka Time Series Tree).

RotationForest([n_estimators, min_group, ...])

A rotation forest (RotF) vector classifier.

Base#

BaseClassifier()

Abstract base class for time series classifiers.

BaseDeepClassifier([batch_size, random_state])

Abstract base class for deep learning time series classifiers.

BaseEarlyClassifier()

Abstract base class for early time series classifiers.