The default behavior for adding and solving with noisemodels has changed from Pastas 1.5. Find more information here

pastas.stats.signatures.high_pulse_count#

high_pulse_count(series, quantile=0.8, rolling_window='7D')[source]#

Average number of times the series exceeds a certain threshold per year.

Parameters
  • series (pandas.Series) – Pandas Series with DatetimeIndex and head values.

  • quantile (float, optional) – Quantile used as a threshold.

  • rolling_window (str, optional) – Rolling window to use for smoothing the time series. Default is 7 days. Set to None to disable. See the pandas documentation for more information.

Returns

Average number of times the series exceeds a certain threshold per year.

Return type

float

Notes

Number of times during which the head exceeds a certain threshold. The threshold is defined as the 80th percentile of non-exceedance.

Warning

This method is sensitive to measurement noise, e.g., every change is sign in the differences is counted as a pulse. Therefore, it is recommended to smooth the time series first (which is also the default).