vinecopulas.marginals

Created on Thu Feb 22 17:13:20 2024

Functions

best_fit_distributiondiscrete(data[, bound, criterion])	Fits the best discrete distribution to data.
best_fit_distribution(data[, criterion, dists])	Fits the best continuous distribution to data.
pseudodata(data)	Compute the pseudo-observations for the given data (transforms data to standard uniform margins)
pseudodiscr(xcdf, xpmf)	Compute the pseudo-observations for the given variable that is discrete.

Module Contents

vinecopulas.marginals.best_fit_distributiondiscrete(data, bound=False, criterion='BIC')

Fits the best discrete distribution to data.

Arguments:

data : The data which has to be fit as a 1-d numpy array.

bounds : whether the data is bounded

criterion : Either BIC, AIC or ML is used to choose the best distribution

Returns:

bestdist : the best distribution and its parameters.

vinecopulas.marginals.best_fit_distribution(data, criterion='BIC', dists=[])

Fits the best continuous distribution to data.

Arguments:

data : The data which has to be fit as a 1-d numpy array.

criterion : Either BIC, AIC or ML is used to choose the best distribution

dists : Specify specific distributions if only specific distributions need to be tested, provided as a list.

Returns:

bestdist : the best distribution and its parameters.

vinecopulas.marginals.pseudodata(data)

Compute the pseudo-observations for the given data (transforms data to standard uniform margins)

Arguments:

data : The data which has to be converted into pseudo data, provided as a numpy array where each column contains a separate variable (eg. x1,x2,…,xn)

Returns:

u : Pseudo data, provided as a numpy array where each column contains a separate variable (eg. u1,u2,…,un)

vinecopulas.marginals.pseudodiscr(xcdf, xpmf)

Compute the pseudo-observations for the given variable that is discrete.

Arguments:

xcdf : The cumulative distribution function of the variable, calculated based on the best fit discrete distribution, provided as a 1-d numpy array.

xpmf : The probability mass function of the variable, calculated based on the best fit discrete distribution, provided as a 1-d numpy array.

Returns:

ui : Pseudo data of a given variable provided as a 1-d numpy array.