copul is a package designed for mathematical computation with and visualization of bivariate copula families.
Install the copul library using pip.
pip install copul
A guide and documentation is available at https://copul.readthedocs.io/.
The copul
package covers implementations of the following copula families:
-
Archimedean copula families: The 22 Archimedean copula families from the book "Nelsen - An Introduction to Copulas" including
- Clayton
- Gumbel-Hougaard
- Frank
- Joe
- Ali-Mikhail-Haq
- etc.
-
Extreme-value copulas families:
- BB5
- Cuadras-Augé
- Galambos
- Gumbel
- Husler-Reiss
- Joe
- Marshall-Olkin
- tEV
- Tawn
-
Elliptical copula families:
- Gaussian
- Student's t
- Laplace.
-
Other copula families:
- Farlie-Gumbel-Morgenstern
- Fréchet
- Mardia
- Plackett
- Raftery
Furthermore, the package provides the following copulas:
- Independence copula
- Lower and upper Fréchet bounds
- Checkerboard copulas
The following properties are available for the above copula families and copulas if they exist and are known:
-
cdf
: Cumulative distribution function -
pdf
: Probability density function -
cond_distr_1
,cond_distr_2
: Conditional distribution functions -
lambda_L
,lambda_U
: Lower and upper tail dependence coefficients -
rho
,tau
,xi
: Spearman's rho, Kendall's tau, and Chatterjee's xi -
generator
,inv_generator
: Generator and inverse generator for Archimedean copula families -
pickands
: Pickands dependence functions for extreme-value copula families
The following methods are available for the above copula families and copulas:
-
rvs
: Generate random samples from the copula -
scatter_plot
: Generate a scatter plot of the copula -
plot_cdf
: Visualize the cumulative distribution function -
plot_pdf
: Visualize the probability density function -
plot_rank_correlations
: Visualize Spearman's rho, Kendall's tau, and Chatterjee's xi -
plot_generator
: Visualize the generator function -
plot_pickands
: Visualize the Pickands dependence function