New plotting functions: circular boxplots with plot_joint_box() and circular
histograms with plot_circ_hist().
New functions to assess the goodness of fit of copulas.
cramer_vonmises() calculates the Cramér-von-Mises criterion and the
associated p-value to compare a parametric copula to the empirical copula
of the data.wasserstein() calculates the Wasserstein distance between two copula PDFs
or between a copula PDF and pseudo-observations.New functions for density, distribution, quantiles and random number generation of linear mixture distributions with an arbitrary number of components.
dweibullmix(), pweibullmix(), qweibullmix(), and rweibullmix().dgammamix(), pgammamix(), qgammamix(), and rgammamix().dnormmix(), pnormmix(), qnormmix(), and rnormmix().dlnormmix(), plnormmix(), qlnormmix(), and rlnormmix().New functions for density, distribution, quantiles and random number generation of
von Mises mixture distributions with an arbitrary number of components:
dvonmisesmix(), pvonmisesmix(), qvonmisesmix(), and rvonmisesmix().
qvonmisesmix() replaces the function qmixedvonmises(), which is deprecated.
New functions to calculate density and distribution and to generate samples
of bivariate joint distributions that are defined in terms of two marginal
distributions and a copula: djoint(), pjoint(), and rjoint().
New function to obtain trajectories from x- and y-coordinates, traj_get()
New checking routines that check every argument of every function for validity.
The plotting functions have new names and added and changed functionalities. The old functions are deprecated.
traj_plot() is now plot_track().
It can now plot multiple trajectories and also separate x- and
y-coordinates.scat_plot() is now plot_joint_scat().
It can take custom kernel density estimates for the margins as
input and the plots have an improved appearance. It can now take as input
trajectories as well as separate angles and step lengthscop_scat_plot() is now plot_cop_scat().
It has new input arguments but with the same functionalities.circ_plot is now plot_joint_circ().
It can now take as input
trajectories as well as separate angles and step lengths.cop_plot is now plot_cop_surf().make_traj() (deprecated) is now called traj_sim().
It can now take as argument
directly the output of fit_steplength() or fit_angle().The performance of many density, distribution, quantile and random number
generating functions is now improved. Especially for copulas of classes
cyl_vonmises and cyl_rect_combine and cyl_cubsec.
The names of the arguments of opt_lin_bw() and opt_circ_bw(), have changed
and are now more intuitive.
mi_cyl() gives now exactly 1 for copulas with perfect correlation if
normalize=TRUE and symmetrize=TRUE are selected.
rdens() (random number generation from kernel density estimates)
now works correctly with circular density estimates.
fit_angle() and fit_steplength() now work correctly with all possible
parametric distributions and return the correct distribution names.
When you plot a copula using plot() and provide additional arguments that go
to the generic plotting function base::plot() there was an error that is now
fixed.
First public release.