direpack.dicomo.dicomo.dicomo
- class dicomo(est='arithmetic', mode='mom', center='mean')[source]
The dicomo class implements (co)-moment statistics, covering both clasical product-moment statistics, as well as more recently developed energy statistics. The dicomo class also serves as a plug-in into capi and ppdire. It has been written consistently with ppdire such that it provides a wide range of projection indices based on (co-)moments. Ancillary functions for (co-)moment estimation are in _dicomo_utils.py.
- Parameters
est (str) – mode of estimation. The set of options are ‘arithmetic’ (product-moment) or ‘distance’ (energy statistics)
mode (str) – type of moment. Options include : * ‘mom’: moment * ‘var’: variance * ‘std’: standard deviation * ‘skew’: skewness * ‘kurt’: kurtosis * ‘com’: co-moment * ‘M3’: shortcut for third order co-moment * ‘cov’: covariance * ‘cos’: co-skewness * ‘cok’: co-kurtosis * ‘corr’: correlation, * ‘continuum’: continuum association * ‘mdd’: martingale difference divergence (requires est = ‘distance’) * ‘mdc’: martingale difference correlation (requires est = ‘distance’) * ‘ballcov’: ball covariance (requires installing Ball and uncommenting the import statement)
center (str) – internal centring used in calculation. Options are mean or median.
- Attributes always provided
moment_: The resulting (co-)moment Depending on the options picked, intermediate results are stored as well, such as x_moment_, y_moment_ or co_moment_
Methods
__init__([est, mode, center])fit(x, **kwargs)Fit a dicomo model
get_params([deep])Get parameters for this estimator.
set_params(**params)Set the parameters of this estimator.
Attributes