aeon v1.2.0

See our changelog to view all changes for this release.

Highlights

• aeon v1.2.0 drops support for Python 3.9, Python 3.10-3.13 are the currently supported versions.
• Framework for whole-series anomaly detection has been implemented, and the module has been refactored into anomaly_detection.collection and anomaly_detection.series submodules.
• ProximityForest and ProximityTree classifiers have been improved to support unequal length and multivariate series. The classifiers are also significantly faster.
• A new implementation for the TD-MVDC classifier
• The forecasting module has been enhanced with additional methods for direct and recursive forecasting, as well as new algorithms such as ETS. This module is still in early development, and we expect framework changes in future releases.
• The similarity search module has been significantly reworked to fit the aeon style framework seen in other modules. See the module examples and documentation for more details.
• The start of an experimental sub-package for self supervised learning has been added in aeon.transformations.collection.self_supervised, starting with the TRILITE algorithm.

New Contributors

• @Ahmed-Zahran02 made their first contribution in #2737
• @SomtoOnyekwelu made their first contribution in #2764
• @saadaltohamy made their first contribution in #2644
• @nMaax made their first contribution in #2622
• @AnaghDeshpande made their first contribution in #2546

All Contributors

@aadya940, @adityagh006, @Ahmed-Zahran02, @AnaghDeshpande, @aryanpola, @baraline, @hadifawaz1999, @inclinedadarsh, @Kaustbh, @kavya-r30, @kevinzb56, @lucifer4073, @MatthewMiddlehurst, @nMaax, @patrickzib, @pattplatt, @Ramana-Raja, @saadaltohamy, @SebastianSchmidl, @shinymack, @SomtoOnyekwelu, @steenrotsman, @tanishy7777, @TinaJin0228, @TonyBagnall

Full Changelog: v1.1.0...v1.2.0

aeon v1.1.0

See our changelog to view all changes for this release.

Highlights

• Python 3.13 is now supported and dependency bounds have been raised
• df-list collections now require (n_cases, n_channels, n_timepoints) formatting. Make sure each dataframe in the list has channels as the first dimension and timepoints are the second.
• The ROCKAD anomaly detector has been added
• THe KASBA clusterer has been added
• Lots of documentation improvements and bug fixes

New Contributors

• @vedpawar2254 made their first contribution in #2424
• @adilsonmedronha made their first contribution in #2440
• @dschrempf made their first contribution in #2464
• @inclinedadarsh made their first contribution in #2489
• @shinymack made their first contribution in #2488
• @kevinzb56 made their first contribution in #2487
• @Akhil-Jasson made their first contribution in #2483
• @sumana-2705 made their first contribution in #2419
• @YashviMehta03 made their first contribution in #2494
• @lucifer4073 made their first contribution in #2518
• @Nikitas100 made their first contribution in #2484
• @tanishy7777 made their first contribution in #2613
• @Kaustbh made their first contribution in #2589
• @TinaJin0228 made their first contribution in #2717
• @ayushsingh9720 made their first contribution in #2617
• @HaroonAzamFiza made their first contribution in #2738
• @adityagh006 made their first contribution in #2735
• @Val-2608 made their first contribution in #2640
• @Ramana-Raja made their first contribution in #2704

All Contributors

@aadya940, @adilsonmedronha, @adityagh006, @Akhil-Jasson, @aryanpola, @ayushsingh9720, @chrisholder, @Cyril-Meyer, @dschrempf, @hadifawaz1999, @HaroonAzamFiza, @inclinedadarsh, @itsdivya1309, @Kaustbh, @kavya-r30, @kevinzb56, @lucifer4073, @MatthewMiddlehurst, @Nikitas100, @notaryanramani, @patrickzib, @pattplatt, @Ramana-Raja, @shinymack, @sumana-2705, @tanishy7777, @TinaJin0228, @TonyBagnall, @Val-2608, @vedpawar2254, @YashviMehta03

Full Changelog: v1.0.0...v1.1.0

aeon v1.0.0

First major release of aeon.

See our changelog to view all changes for this release.

Highlights

• Introduction of a new experimental forecasting module, replacing the old module (this is currently sparse, but we will expand on the framework and available estimators in the future).
• Restructuring and expansion of the clustering module, including deep learning clustering.
• Explicit split of the transformation module into series transformers and collection transformers.
• Extended anomaly detection module.
• Completely new testing suite, code coverage now 91%.
• Enhanced visualisation tools.
• Improved documentation and examples.
• Introduction of more multivariate capable deep learning estimators.

New Contributors

• @notaryanramani made their first contribution in #2029
• @danielroberts20 made their first contribution in #2064
• @ferewi made their first contribution in #2091
• @LinGinQiu made their first contribution in #2141
• @georgejl7 made their first contribution in #2173
• @fspinna made their first contribution in #2062
• @emmanuel-ferdman made their first contribution in #2275
• @pattplatt made their first contribution in #2342
• @kavya-r30 made their first contribution in #2364

All Contributors

@aadya940, @alexbanwell1, @aryanpola, @baraline, @chrisholder, @Cyril-Meyer, @danielroberts20, @Datadote, @emmanuel-ferdman, @ferewi, @fspinna, @georgejl7, @griegner, @hadifawaz1999, @IRKnyazev, @itsdivya1309, @kavya-r30, @LinGinQiu, @MatthewMiddlehurst, @Moonzyyy, @notaryanramani, @patrickzib, @pattplatt, @SebastianSchmidl, @Sharathchenna, @TonyBagnall

Full Changelog: v0.11.1...v1.0.0

aeon v0.11.1

See our changelog to view all changes for this release.

This release will be the last 0.X minor release. Other than patches, the next release will be 1.0.0 which contains a large amount of framework changes.

Highlights

• STOMP anomaly detector
• QUANT regressor
• Series search for similarity search module
• Piecewise Linear Approximation (PLA) collection transform

New Contributors

• @wenig made their first contribution in #2031

All Contributors

@baraline, @CodeLionX, @Cyril-Meyer, @dguijo, @IRKnyazev, @MatthewMiddlehurst, @Moonzyyy, @TonyBagnall, @wenig

Full Changelog: v0.11.0...v0.11.1

aeon v0.11.0

See our changelog to view all changes for this release.

This release will be the last 0.X minor release. Other than patches, the next release will be 1.0.0 which contains a large amount of framework changes.

Highlights

• Added Proximity Forest classifier
• Added a composable ensemble for classification and regression
• New deep learning networks in AEBiGRUNetwork, AEDRNNNetwork and AEAttentionBiGRUNetwork
• Lots of bug and documentation fixes

New Contributors

• @PatriceJada made their first contribution in #1619
• @IRKnyazev made their first contribution in #1842
• @Cyril-Meyer made their first contribution in #1851
• @Datadote made their first contribution in #1873
• @phershbe made their first contribution in #1915
• @Sharathchenna made their first contribution in #1953
• @aryanpola made their first contribution in #1949

All Contributors

@aadya940, @aryanpola, @baraline, @chrisholder, @CodeLionX, @Cyril-Meyer, @Datadote, @dguijo, @ghost, @hadifawaz1999, @harshithasudhakar, @IRKnyazev, @itsdivya1309, @MatthewMiddlehurst, @Moonzyyy, @PatriceJada, @phershbe, @Sharathchenna, @TonyBagnall

Full Changelog: v0.10.0...v0.11.0

aeon v0.10.0

See our changelog to view all changes for this release.

Highlights

• Dropped support for Python 3.8
• Announcement of deprecations for v1.0.0, includes removal and rework of the current forecasting and transformation frameworks
• New anomaly detection methods: DWT-MLEAD, K-Means, and an adapter for PyOD
• New data loaders for anomaly detection and segmentation benchmark archives
• New Proximity Tree classifier using aeon distances, Proximity Forest to follow.*
• New feature-based and dummy clusterers
• New FLUSS and BinSeg segmenters
• Addition of visualisation tools for shapelets

New Contributors

• @futuer-szd made their first contribution in #1640
• @Moonzyyy made their first contribution in #1663
• @adm-unl made their first contribution in #1707
• @Abhash297 made their first contribution in #1736
• @ermshaua made their first contribution in #1755

All Contributors

@aadya940, @Abhash297, @adm-unl, @baraline, @chrisholder, @CodeLionX, @ermshaua, @futuer-szd, @hadifawaz1999, @itsdivya1309, @MatthewMiddlehurst, @Moonzyyy, @patrickzib, @TonyBagnall, @zy18811

Full Changelog: v0.9.0...v0.10.0

aeon v0.9.0

See our changelog to view all changes for this release.

Highlights

• New anomaly detection base class and MERLIN anomaly detector
• Implementation for the RSAST classifier
• Fixes for pipelines and new clustering and collection transformer pipeline classes
• Improved model loading ability for deep learning modules
• Large number of deprecations as we port transformations over to a new interface

New Contributors

• @nileenagp made their first contribution in #1472
• @jasonmokk made their first contribution in #1515
• @nirojasva made their first contribution in #1383
• @maxwell1503 made their first contribution in #1563

All Contributors

@aadya940, @AnonymousCodes911, @baraline, @chrisholder, @CodeLionX, @hadifawaz1999, @itsdivya1309, @jasonmokk, @MatthewMiddlehurst, @maxwell1503, @nileenagp, @nirojasva, @TonyBagnall

Full Changelog: v0.8.1...v0.9.0

aeon v0.8.1

See our changelog to view all changes for this release.

Highlights

• Mostly bug fixes, documentation improvements and new deprecations
• Subgradient barycentre averaging is available for clustering

New Contributors

• @wayneadams made their first contribution in #1396
• @griegner made their first contribution in #1458
• @RishavKumarSinha made their first contribution in #1462

All Contributors

@aadya940, @AnonymousCodes911, @chrisholder, @CodeLionX, @hadifawaz1999, @harshithasudhakar, @itsdivya1309, @MatthewMiddlehurst, @TonyBagnall, @tvilight4, @vNtzYy

Full Changelog: v0.8.0...v0.8.1

aeon v0.8.0

See our changelog to view all changes for this release.

Highlights

• Deprecations from 0.7.0 have been removed, see the package deprecation sections for more details
• typing-extensions has been added as a core dependency
• The following deep learners are now available for regression: IndividualLITERegressor, LITETimeRegressor and EncoderRegressor
• The HydraRegressor and MultiRocketHydraRegressor algorithms have been implemented for regression module
• A wrapper for the tslearn LearningShapelets classifier has been added
• Support for unequal length in pairwise distance calculation for the SBD and MSM distances is now available, this will be expanded to other distances in time

New Contributors

• @harshithasudhakar made their first contribution in #1299
• @tvilight4 made their first contribution in #1213
• @vNtzYy made their first contribution in #1329

All Contributors

@aadya940, @AnonymousCodes911, @chrisholder, @CodeLionX, @hadifawaz1999, @harshithasudhakar, @itsdivya1309, @MatthewMiddlehurst, @TonyBagnall, @tvilight4, @vNtzYy

Full Changelog: v0.7.1...v0.8.0

aeon v0.7.1

See our changelog to view all changes for this release.

Highlights

• Adds the RIST, Hydra, MR-Hydra and QUANT classifiers
• Adds the above for regression also, as well as MLPRegressor, RDSTRegressor and simple feature-based regressors
• Adds the SBD distance
• Website updates and improvements, including advertisement for GSoC 2024!

New Contributors

• @Raya679 made their first contribution in #1163
• @itsdivya1309 made their first contribution in #1146
• @nimanzik made their first contribution in #1192
• @Vedant222 made their first contribution in #1203
• @aadya940 made their first contribution in #1215
• @AnonymousCodes911 made their first contribution in #1245
• @CodeLionX made their first contribution in #1236

All Contributors

@aadya940, @AnonymousCodes911, @baraline, @chrisholder, @CodeLionX, @hadifawaz1999, @itsdivya1309, @MatthewMiddlehurst, @nimanzik, @Raya679, @TonyBagnall, @Vedant222

Full Changelog: v0.7.0...v0.7.1