Accelerometer

What's Changed
* Fix GENEA timezone issue and DST correction by angerhang in https://github.com/OxWearables/biobankAccelerometerAnalysis/pull/211
* Fix wrong model URLs: `willetts-10Feb2022` and `doherty-10Feb2022` were both incorrectly pointing to `walmsley-10Feb2022`
c27d25d3ce05c51ee10840f70ff0c2d19a911e82
* License amendment 95715078021004d89a2aa94afc96b83c4979e6d4
* Improvements in documentation.

What's Changed
Bug fixes
* critical fix to java path when running on Windows 68257a7dd18ea3030cd25085df84ef940279e862
* fix plot bug 573491ff1c623a99540e4f82be5b1c588aa9c364

Improvements
* speed up missing data imputation 47dc1fcc7a49395621dcaedb23755d6d84b457f2

What's Changed
* Fixed a bug during calibration when there are constant variables e80506b61c9d8522ab1efccdeda83e1baba0e3b1
* Improved documentation: Included a note on device axis orientation in the widely used UK Biobank accelerometer dataset.
* License change

What's Changed
* Fix a bug in pandas inferred freq by chanshing in https://github.com/activityMonitoring/biobankAccelerometerAnalysis/pull/194
* Improve documentation
* Added a data dictionary section: https://biobankaccanalysis.readthedocs.io/en/latest/datadict.html
* Added a quality control section: https://biobankaccanalysis.readthedocs.io/en/latest/usage.html#quality-control

What's Changed
* Improve classification module: Add support for cross-validation and custom `max_depth` and `min_samples_leaf` (useful to reduce model size). Improve reporting by outputting a JSON file with model metrics.
* Update trained models (much lighter now) and model URLs.
* Fix `writeCmds` when paths contain spaces

What's Changed

- The Capture-24 dataset was revised for annotation errors and ML models were retrained accordingly. Differences in the ML classification outputs might be expected.
- Major change in the random forest models to be based on [imbalanced-learn](https://imbalanced-learn.org/) package to replace the monkeypatch approach employed so far c14c1a35a4a2b4a2dc5a8ceae592a87e2438bff2
- Fix spectral entropy normalization feature 8abdc32c11d4e074584b91284be0c06b06fa908a
- No longer compute/use fft3d features (expensive and not really useful). Also reduce number of fft features from 12 to 10 aca88ca4cae23ec9d160697c81f9c2314fe5375e
- Github Actions to no longer run classification as the new models are a bit too large.
- Minor refactoring: summary variable name changes (`wear` -> `wearTime`, `CutPoint*` -> `cutPoint*`, etc.)
- Other small improvements.

Due to the changes above, older activity classification models are no longer supported (you can still use them reverting to older releases).