StarVine provides tools to construct canonical and regular-vines (C-vines, and R-vines).
StarVine can also be used as a standalone copula fitting tool for bivariate modeling.
The following bivariate copula families are built-in:
- Elliptic
- Gauss
- Student-t
- Archimedean
- Frank
- Gumbel
- Clayton
- Marshall-Olkin
- Mixture (Work in progress)
All copula orientations are implemented.
StarVine can also be used to fit and sample C-Vine:
Install the package:
python setup.py install
For a developer install:
python setup.py develop --user
Requires:
- A fortran compiler
- numpy : basic linear algebra operations
- scipy : additional numerical operations
- networkx : vine graph structure
- h5py : file I/O
- emcee : MCMC
- seaborn : plotting
- matplotlib : basic plotting, required by seaborn
- pandas
Optional:
- statsmodels: Generalized method of moments fitting
- mpi4py : parallel MCMC
Documentation can be found online at: wgurecky.github.io/starvine.
The docs can be manually built using Doxygen:
doxygen Doxyfile.in
StarVine is distributed under the BSD 3 clause license.
A copy of the license should have been distributed with the StarVine source code. If not, see opensource.org