forecasting

functime supports both individual forecasters and forecasters with automated lags / hyperparameter tuning. Auto-forecasters uses FLAML to optimize both hyperparameters and number of lagged dependent variables. FLAML is a SOTA library for automated hyperparameter tuning using the CFO (Frugal Optimization for Cost-related Hyperparamters¹) algorithm. All individual forecasters (e.g. lasso / xgboost) and automated forecasters (e.g. auto_lasso and auto_xgboost) implement the following API.

forecaster

Autoregressive forecaster.

Parameters:

Name	Type	Description	Default
freq	str	Offset alias supported by Polars.	required
lags	int	Number of lagged target variables.	required
max_horizons	Optional[int]	Maximum number of horizons to predict directly. Only applied if strategy equals "direct" or "ensemble".	None
strategy	Optional[str]	Forecasting strategy. Currently supports "recursive", "direct", and "ensemble" of both recursive and direct strategies.	None
target_transform	Optional[Transformer]	functime transformer to apply to y before fit. The transform is inverted at predict time.	None
feature_transform	Optional[Transformer]	functime transformer to apply to X before fit and predict.	None
**kwargs	Mapping[str, Any]	Additional keyword arguments passed into underlying sklearn-compatible regressor.	{}

auto_forecaster

Forecaster with automated hyperparameter tuning and lags selection.

Parameters:

Name	Type	Description	Default
freq	str	Offset alias as dictated.	required
min_lags	int	Minimum number of lagged target values.	3
max_lags	int	Maximum number of lagged target values.	12
max_horizons	Optional[int]	Maximum number of horizons to predict directly. Only applied if strategy equals "direct" or "ensemble".	None
strategy	Optional[str]	Forecasting strategy. Currently supports "recursive", "direct", and "ensemble" of both recursive and direct strategies.	None
test_size	int	Number of lags.	1
step_size	int	Step size between backtest windows.	1
n_splits	int	Number of backtest splits.	5
time_budget	int	Maximum time budgeted to train each forecaster per window and set of hyperparameters.	5
search_space	Optional[dict]	Equivalent to config in FLAML	None
points_to_evaluate	Optional[dict]	Equivalent to points_to_evaluate in FLAML	None
num_samples	int	Number of hyper-parameter sets to test. -1 means unlimited (until time_budget is exhausted.)	-1
target_transform	Optional[Transformer]	functime transformer to apply to y before fit. The transform is inverted at predict time.	None
feature_transform	Optional[Transformer]	functime transformer to apply to X before fit and predict.	None
**kwargs	Mapping[str, Any]	Additional keyword arguments passed into underlying sklearn-compatible regressor.	{}

functime also has the following benchmark models implemented as pure Polars queries.

naive

Naive forecaster.

Parameters:

Name	Type	Description	Default
freq	str	Offset alias supported by Polars.	required

snaive

Seasonal naive forecaster.

Parameters:

Name	Type	Description	Default
freq	str	Offset alias supported by Polars.	required
sp	int	Seasonal periods.	required

---

1. https://arxiv.org/abs/2005.01571 ↩