search

time-series-machine-learning / tsml-eval

Evaluation tools for time series machine learning algorithms.
https://tsml-eval.readthedocs.io/
BSD 3-Clause "New" or "Revised" License
32 stars 14 forks source link

Run all univariate capable `aeon` classifiers on extended univariate TSC datasets #3

Open TonyBagnall opened 2 years ago

TonyBagnall commented 2 years ago

we are extending to 112 to around 150, including 20+ new problems and equal length versions of the current unequal length data. To do:

  1. Finalise list and check all data load from disk and TSC.com
  2. Run HC2 components, including train files
  3. HC2 proper, no train files
  4. Compose HC2, verify same results as (2)
  5. Other classifiers (form list and post here)
  6. Deep learning classifiers
MatthewMiddlehurst commented 1 year ago

Currently running (2) on the 112 datasets. Can run the new datasets when they are ready.

TonyBagnall commented 1 year ago

if you focus on the 112, I can do the new ones. Are you running with sktime default parameter?

MatthewMiddlehurst commented 1 year ago

using set_classifier, which does deviate I'm pretty sure

TonyBagnall commented 1 year ago

yeah we should check that, I should have thought about that!

MatthewMiddlehurst commented 1 year ago

good to compare both either way