Aeon Toolkit (aeon)

v1.3.0 is now available on PyPI. Changelog lnkd.in/eAZKzGR4. Thanks to the combined work of our contributors and GSoC students, we now have a lot more to offer in our forecasting module, and lot of other new contributions for other modules as well

The notebook accompanying Tony's blog is github.com/aeon-toolkit/a…

New blog post on forecasting horizons by @AHIsmailFawaz @tony_bagnall @aeon.toolkit/forecasting-horizons-da426b7a6c24" target="_blank" rel="nofollow noopener">medium.com/@aeon.toolkit/…

We are going to start writing articles on medium on the toolkit functionality, first post is on our new forecasting module @aeon.toolkit/blazingly-fast-arima-with-aeon-9796d7015cdd" target="_blank" rel="nofollow noopener">medium.com/@aeon.toolkit/…

My first medium post for aeon. Its about our new bespoke implementation of ARIMA that uses a numba version of Nelder-Mead to fit. Its 10x faster than the best alternative @aeon.toolkit/blazingly-fast-arima-with-aeon-9796d7015cdd" target="_blank" rel="nofollow noopener">medium.com/@aeon.toolkit/…

the deadline for AALTD 2025 is approaching, and if you are thinking of submitting please submit an abstract by 7th June. You can change it after, or not actually submit, it just helps us assign reviewers ecml-aaltd.github.io/aaltd2025/

New Paper: "Rock the KASBA: Blazingly Fast and Accurate Time Series Clustering". arxiv.org/abs/2411.17838 State of the art clustering of all UCR datasets with KESBA in under two hours on a laptop. Soon to be in @aeon_toolkit

aeon has hit 1000 stars on github. Steady organic growth was our plan and it seems to be working. We now have 15 core developers of eight different nationalities and over 70 contributors from across the world. We have big plans for version 1.0, coming soon

Our paper in JMLR Machine Learning Open Source Software track is finally available. It is a bit out of date now, the toolkit is adapting and growing, but still nice to get the paper. jmlr.org/papers/v25/23-…

We are welcoming another new core developer to aeon, welcome to the team Divya! 🥳

We are glad to welcome Aadya as a new core developer to aeon! Let's wish him a warm welcome! 🥳

the aeon roadshow continued with a tutorial at @ECMLPKDD in Vilnius. We had a blast, and have exciting plans for the next year. Room was packed and we think it went down well. Slides and code here github.com/aeon-toolkit/a…

aeon core developers and friends were at KDD in Barcelona presenting a hands on tutorial on time series classification and regression with aeon. The room was packed, standing room only with over 100 in attendance. github.com/aeon-toolkit/a…

The last minor release before aeon 1.0 is out! It includes new deep learning networks, composable estimators, Proximity Forest classifier, and lots of bugfixes and documentation improvements. Thanks to the 7 new contributors that joined us this patch! 🙏 Full changelog on GitHub

This week we released aeon v0.10, with 15 contributors since last release, including 5 new to aeon. We also announced our plan to deprecate legacy code in version 1.0, coming this Summer. This includes the entire forecasting module. New forecasting implementations coming soon.

Divya Tiwari is a Google Summer of Code intern. She is working on implementing the Proximity Forest time series classifier by Geoff Webb's group. Her blog is here. @Divya2003/gsoc24-with-aeon-numfocus-cc976b9b2fc5" target="_blank" rel="nofollow noopener">medium.com/@Divya2003/gso… @AHIsmailFawaz @giwebb

Aadya Chinubhai is a Google Summer of Code intern. He is working on implementing the best deep learning clusterers from here link.springer.com/article/10.100… blog posts here @aadyachinubhai" target="_blank" rel="nofollow noopener">medium.com/@aadyachinubhai @AHIsmailFawaz

aeon are running 11 projects over the summer, Google Summer of Code, interns and MSc students. Its all very exciting and start of phase 2 of our development

Last week we released aeon v0.9 which introduces the foundation of the new anomaly detection module, some fixes to pipelines, improved loading capacities for deep learning networks and deprecations to move transformations over to a new interface in 0.10. Thx to all contributiors!