incline.loess_trend

incline.loess_trend(df, column_value='value', time_column=None, frac=0.3, degree=1, robust=True, **kwargs)[source]

Estimate the trend with LOESS.

robust=True reweights according to the residuals, which makes the fit data-dependent; standard errors then come from the bootstrap. With robust=False the smoother is linear and they are exact.

Parameters:

  • df (pd.DataFrame) – Time series data.

  • column_value (str) – Column holding the values.

  • time_column (str | None) – Numeric time column.

  • frac (float) – Fraction of the sample in each local regression.

  • degree (int) – Degree of the local polynomial.

  • robust (bool) – Whether to run robustifying iterations.

  • **kwargs (Any) – Uncertainty options; see estimate().

Returns:

The input frame plus the estimate columns.

Return type:

pd.DataFrame