0.6.1

0.6.1 ✨✨✨

• CenterSplit transformer added
• separated function for best model selection
• improved graphing when forecast_length == 1
• quieter Prophet
• added PreprocessingRegression model
• more KalmanStateSpace models
• assorted bug fixes
• added AutoTS diagnose_params

0.6.0

0.6.0 ♨️♨️♨️

• import_best_model and fit_data
• added canberra distance to SeasonalityMotif and MetricMotif
• DatepartRegressionTransformer now handles NaN in input data
• new DatepartRegression and SeasonalityMotif based imputers
• LevelShiftMagic transformer added
• adjusted automatic resampling to be performed only when necessary (to avoid filling NA with 0 bug and aggfunc='sum')
• fixed a bug where if a full validation round fails, best model selection fails
• macro_micro to LocalLinearTrend

0.5.8

0.5.8 🌷🌷🌷

• added plot_validations
• updated pytorch forecasting for 1.0.0 version
• plot_grid to prediction object
• breaking change: plot_per_series_smape switched to the more accurately described plot_per_series_mape
• various bug fixes such as 'TotalRuntime' missing

0.5.7

0.5.7 🛂🛂🛂

• slight changes to holiday_flag to allow list in some cases
• DatepartRegressionTransformer now accepts holiday country input as regressor
• added RegressionFilter
• changed bounded behavior of AlignLastValue
• small bug fixes

0.5.6

0.5.6 🌌 🌌 🌌

• fixes for annoying things broken by pandas 2.0, without them giving deprecation warnings
• also for gluonts who also love breaking their API
• ensembling tuning

0.5.5

0.5.5 🌌 🌌 🌌

• updated metrics
• Cassandra bug fixes
• fail_on_forecast_nan now also checks for infinity, and is faster

0.5.4

0.5.4 🌏 🌏 🌏

• Cassandra seasonality, holiday detection, and other bug fixes
• attempted to improve the KalmanStateSpace and KalmanSmoothing methods
• validation indexes now generated in a standalone function
• create_regressor bug fix
• added MLEnsemble
• fix a bug where common_fourier failed on short length forecasts
• added some additional options to ARDL
• refactored metric df eval
• refactored parts of AutoTS for templates and validations
• added ewmae metric
• non Horizontal ensembles are now also constructed after validations and vals are rerun

0.5.3

0.5.3 🌍 🌍 🌍

• robustness changes to generate_score_per_series for horizontal ensembles
• added generation_timeout to allow stopping based on time
• added ability to specify anomaly model (ie 'zscore') to AnomalyDetector.get_new_params(method='zscore')
• fixed a bug in Cassandra with future_impacts and future_regressor

0.5.2

0.5.2 🌎 🌎 🌎

• allow transformer_max_depth == 0
• linear_tree option to LightGBM models
• additional linear model options to Cassandra

0.5.1

0.5.1 🧙‍♂️ 🧙‍♂️ 🧙‍♂️

• add LocalLinearTrend transformer
• improved ScipyFilter focusing on Butter and Savitzky–Golay filters
• update to AutoTS().back_forecast including breaking change of column arg renamed to series
• add KalmanSmoothing transformer
• add KalmanStateSpace model (Kalman Filter + 'any' state space models)
• AlignLastValue no longer applied to upper/lower forecast bounds
• modified regression impact in HolidayTransformer to weighted least squares, moved existing to 'datepart_regression'
• added Cassandra model
• bug fix for Categorical dateparts with 1 starts
• updated load_daily to Wikimedia page views
• added dwae metric
• added MetricMotif model
• added 'common_fourier' datepart method
• added SeasonalMotif model
• added 'seasonal' validation
• bug fix for cffilter with univariate input