aeon
aeon copied to clipboard
aeon-toolkit
A toolkit for machine learning from time series
got a bit bigger than I expected, so will put up as draft and note changes Fixes #2135: changes parameter name Fixes #2134; requires weights passed as ndarray Removes unused...
#### Reference Issues/PRs #2130 #### What does this implement/fix? Explain your changes. This PR implements the Elastic-SOM clusterer. SOM is a neural network that iteratively updates weights between input neurons...
#### Reference Issues/PRs Part1 to Fixes #2071 #### What does this implement/fix? Explain your changes. Remove distance argument from RDST, uniformize dtypes, use of strides (view of the original array),...
coverage for BOSS, more to do
sets MUSE capability:univariate to False, and adds a test to improve coverage.
draft for updating projects and including list of completed projects. WIP, I need to check people are ok with this and see about links etc
### Describe the feature or idea you want to propose FAILED aeon/distances/tests/test_distances.py::test_distances[dist16] - assert False + where False = isinstance(0, float) = 1 failed, 8698 passed, 11 skipped, 52 xfailed,...
fixes #2143 adds testing coverage for incorrect uses, too many dimensions and histogram intersection. Not sure whether the numba.Dict test needs to be separate and excluded from test without numba.
### Describe the bug found when writing coverage testing, calls to histogram_intersection in TemporalDictionaryEnsemble throw an error due to the use of np.min instead of np.minimum, probably some numpy change...
improves data coverage for writers. This removes the write to tsf format functionality, because it makes assumptions about how we model a forecasting data. ATM that is not decided, and...
Metadata
Owner
Metadata
A toolkit for machine learning from time series