Research Fellow at the University of Southampton. Interested in machine learning with a focus on time series.
Norwich, United Kingdom
Joined September 2019
- Tweets19
- Following31
- Followers40
- Likes30
Matthew Middlehurst retweeted
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/aeon…
Matthew Middlehurst retweeted
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/aeon…
Matthew Middlehurst retweeted
I am helping to give a hands on tutorial on time series classification and regression at KDD (@kdd_news) on Monday 26th aeon-tutorials.github.io/KDD… with @giwebb @gforestier @AHIsmailFawaz @M_Middlehurst also presenting, examples all using @aeon_toolkit. Should be fun!
Matthew Middlehurst retweeted
Tutorial An Introduction to Machine Learning from Time Series event located at the @ECMLPKDD 2024. Learn more about ML and temporal data! @M_Middlehurst @gforestier @ermshaua
Matthew Middlehurst retweeted
We just released aeon v0.7 with Python 3.12 support, thanks to all contributors !
For more information and highlights, check : github.com/aeon-toolkit/aeon…
Full changelog available here : aeon-toolkit.org/en/latest/c…
Matthew Middlehurst retweeted
I am currently advertising a PhD, come join the BagLab :) Students from UK, EU27 and EU Horizon countries are eligible for full funding. Will involve working with the @aeon_toolkit community findaphd.com/phds/project/ma…
Matthew Middlehurst retweeted
clustering time series? Using DTW? You might want to read this link.springer.com/article/10… A new paper using aeon for clustering with elastic distances, just in time for our @pydataamsterdam talk :) Chris Holder @M_Middlehurst and @tony_bagnall
Matthew Middlehurst retweeted
the 8th Workshop on Advanced Analytics and Learning on Temporal Data will take place at
@ECMLPKDD
2023. Deadline: June 12, 2023, aeon will be supporting details to follow ecml-aaltd.github.io/aaltd20…
@tony_bagnall @rtavenar @heerme
Matthew Middlehurst retweeted
v0.2.0 of the aeon time series machine learning toolkit has been released on PyPi. Change log here aeon-toolkit.org/en/latest/c… highlights: support python 3.11; new regressors; new dilated shapelet transform; distance functions now much faster; getting started page reworked
Matthew Middlehurst retweeted
The first full release of aeon is available with aeon==0.1.0! 🚀 It comes with pandas==2.0.0 support to speed up forecasters and updated implementations for other modules. Check out the full release notes on our changelog. You can install aeon with "pip install aeon".
Matthew Middlehurst retweeted
aeon toolkit version 0.1.0 is on pypi, follow us here for updates nitter.gq/aeon_toolkit
Matthew Middlehurst retweeted
You can now use aeon-toolkit to use Deep Learning Models for Time Series Classification including InceptionTime, CNN etc. with our hand-crafted filters !
github.com/aeon-toolkit/aeon…
aeon-toolkit : aeon-toolkit.org
custom filters : maxime-devanne.com/pages/HCC…
![[ENH] For classification: Adding InceptionTime, CNN and ResNet along with other classifiers and...](https://pbs.twimg.com/card_img/2114193240509251584/2K2xyD_-?format=jpg&name=800x419)
I would like to share that my paper on "Deep Learning For Time Series Classification Using New Hand-Crafted Convolution Filters" was accepted at the 2022 IEEE International Conference on Big Data done with @MaximeDevanne, @jjmweber and @gforestier.
#IEEEBigData2022
Version 0.1.0 of the aeon toolkit is out! This is only the start though, more improvements and time series algorithms to come! Follow @aeon_toolkit for more updates.
Matthew Middlehurst retweeted
our paper on HIVE-COTEv2.0, a new state of the art for time series classification, is out and open access link.springer.com/article/10… @sktime_toolbox @MLJ_Social @M_Middlehurst @JasonLines @jammylarge @M_J_Flynn @_Groshh_
Matthew Middlehurst retweeted
"HIVE-COTE 2.0: a new meta ensemble for time series classification" arxiv.org/abs/2104.07551
Gave my Q&A on the Temporal Dictionary Ensemble (TDE) for time series classification earlier at #ecmlpkdd2020 and will be doing another later today. Thanks to those who asked questions and attended the first one!
Matthew Middlehurst retweeted
Cool work by @tony_bagnall @jammylarge integrating catch22 features into new time-series classifiers "significantly more accurate on the UCR archive than any other classifier we are aware of and represents a new state-of-the-art"
arxiv.org/abs/2008.09172
