Seasonal decompose features

`polars_ts.decomposition.seasonal_decompose_features`

Perform seasonal decomposition on a time series and compute additional features.

Parameters:

Name	Type	Description	Default
`df`	`DataFrame`	A Polars DataFrame containing the time series data with the following columns: `id_col`: Identifier for each time series (e.g., product, region, etc.) `time_col`: Timestamp column indicating the time of each observation. `target_col`: The column containing the values to be decomposed.	required
`ts_freq`	`str`	The frequency of the time series. For example, `24` could represent daily seasonality in an hourly time series.	required
`seasonal_freqs`	`Optional[List[int]]`	A list of seasonal frequencies to use for the MSTL mode. This is required only if `mode='mstl'`. Defaults to None.	`None`
`mode`	`Literal['simple', 'mstl']`	The decomposition mode, one of {'simple', 'mstl'}. Defaults to 'simple'.	`'simple'`
`id_col`	`str`	The name of the column that identifies each individual time series within the DataFrame. Defaults to `unique_id`.	`'unique_id'`
`time_col`	`str`	The name of the column that contains the time or datetime information for each observation. Defaults to `ds`.	`'ds'`
`target_col`	`str`	The name of the column containing the time series data to be decomposed. Defaults to `y`.	`'y'`

Returns:

Type Description

DataFrame

A DataFrame containing the following features for each unique id_col:

trend_strength: A measure of how strong the trend component is in the time series (value between 0 and 1).
seasonal_strength: A measure of how strong the seasonal component is in the time series (value between 0 and 1).
resid_var: The ratio of the standard deviation of the residuals to the mean of the target variable (a measure of residual variance).