miscΒΆ
Auxiliary functions with miscellaneous use.
FunctionsΒΆ
|
Calculate a Kernel Density Estimate (KDE) distribution for a given dataset. |
|
Get maximum or minimum value of the set with all values of all provided arrays. |
|
Round a number to the nearest (up or down) multiple of a given factor. |
Module ContentsΒΆ
- ensemblify.utils.misc.kde(data, weights=None)ΒΆ
Calculate a Kernel Density Estimate (KDE) distribution for a given dataset.
Weights for the given dataset can be provided to alter the contribution of each data point to the KDE distribution.
- Parameters:
data (
np.ndarray) β Dataset to calculate KDE values for.weights (
list, optional) β Array with weights for each data point. Defaults to uniform weights.
- Returns:
- x_coords (np.ndarray):
X axis coordinates corresponding to the calculated kde distribution.
- norm_kde (np.ndarray):
Normalized kde distribution.
- weighted_average (float):
Weighted average of the given dataset.
- weighted_standard_error (float):
Weighted standard error of the calculated average.
- Return type:
tuple[np.ndarray,np.ndarray,float,float]
- Adapted from:
https://github.com/FrPsc/EnsembleLab/blob/main/EnsembleLab.ipynb
- Reference for standard error of weighted average:
https://seismo.berkeley.edu/~kirchner/Toolkits/Toolkit_12.pdf Copyright Β© 2006 Prof. James Kirchner
- ensemblify.utils.misc.get_array_extremum(arrays, get_max=True)ΒΆ
Get maximum or minimum value of the set with all values of all provided arrays.
- Parameters:
arrays (
list[np.ndarray]) β List of arrays to analyze.get_max (
bool, optional) β Whether to get the maximum or minimum (if False) value. Defaults to True.
- Returns:
Maximum or minimum value.
- Return type:
float
- ensemblify.utils.misc.round_to_nearest_multiple(n, factor, up=True)ΒΆ
Round a number to the nearest (up or down) multiple of a given factor.
- Parameters:
n (
int) β Number to round.factor (
int) β The number n will be rounded to a multiple of factor.up (
bool, optional) β Whether to round up or down (if False). Defaults to True.
- Returns:
rounded number.
- Return type:
rounded