pyacs.gts.lib.outliers.find_l1trend module

pyacs.gts.lib.outliers.find_l1trend.find_l1trend(self, lam, threshold, period=None, gap=10, components='NE', plot=False, verbose=False, in_place=False)[source]

Find outliers using L1 trend filtering and residual threshold.

Parameters:
  • lam (float) – Lambda parameter for L1 trend filtering.

  • threshold (float) – Residuals above threshold * (scale) are flagged as outliers.

  • period (list or None, optional) – Period(s) for searching outliers.

  • gap (int or float, optional) – Gap in days to split series (default 10).

  • components (str, optional) – Components for detection (default ‘NE’).

  • plot (bool, optional) – If True, plot filter result and flagged outliers.

  • verbose (bool, optional) – Verbose mode.

  • in_place (bool, optional) – If True, modify current Gts; otherwise return a new Gts.

Returns:

Gts with outliers set (new instance or self if in_place).

Return type:

Gts