Releases: zillow/luminaire
Releases · zillow/luminaire
0.2.3
Bug fixes:
- Data reindexing while imputation fixed at the presence of missing / invalid data
Scoring logic updates:
- Model uncertainty is taken into consideration while making stationarity adjustments while scoring WindowDensityModel
0.2.2
Bug fixes:
False positives due to detrending failure for shorter history time series in WindowDensityModel
0.2.1
- Improved automatic window size detection logic added for the WindowDensityModel
0.2.0
- WindowDensity model improvements for streaming and high-frequency time series
- Full automation in training and scoring the window density model
- Minor version upgrades for package dependencies (more on the way!)
- Bugfixes
0.2.0.dev1
Dev release for v0.2.0.
This release includes the following:
- Improved WindowDensity modeling for streaming use cases.
- Bringing automation in configuring window density model for streaming use cases.
0.1.4
- Minor bug fixes applied to hyperparameter optimization.
- Updates in docs and readme.
0.1.3
- training_start and training_end are removed from required parameters for training the window density model
- doc corrections and updates
0.1.1
This release fixes the bug #52 for ‘T’ and ‘15T’ pandas frequencies for window density model.
Minor documentation updates.
0.1.0
Making Luminaire available as a beta release.
Details:
- Core Luminaire code base
- Documentation:
- CI/CD pipeline workflow for build, release and documents
- Improved code/files organization
Doc updates
Merge pull request #35 from zillow/issue26-slowdown-workflow-gif
docs and version update