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

pastas.stats.signatures.bimodality_coefficient#

bimodality_coefficient(series, normalize=True)[source]#

Bimodality coefficient after Ellison [1987].

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

  • normalize (bool, optional) – normalize the time series to values between zero and one.

Returns

Bimodality coefficient.

Return type

float

Notes

Squared product moment skewness (s) plus one, divided by product moment kurtosis (k):

..math::

b = (s^2 + 1 ) / k

Adapted from the R ‘modes’ package. The higher the bimodality coefficient, the more bimodal the head distribution is, and vice versa.