-
Default argument for
p_select
inhystar_fit()
now shows all four options. The new default isp_select = c("bic", "aic", "aicc", "aiccp")
, and the first option"bic"
is taken when the user doesn't provide the argument. -
Not necessary to provide all regime names and colors, can all be specified separately
-
Added
call
tohystar_sim
andhystar_fit
objects. Call now also printed inprint
method. -
The user-feedback about the progress is now not shown by default, but can be requested by setting the
show_progress
argument inhystar_fit()
toTRUE
. This is becauseknitr
did not handle theappendLF = FALSE
argument (ensuring progress gets printed on the same line each iteration, and not on a new line) of themessage()
-function well (that prints the progress-feedback); as a result, the print ofhystar_fit()
in my example on github and on the website was veeery long.
-
Added a welcome message with a logo and first directions for help.
-
Added user-feedback about the progress of the estimation when running
hystar_fit()
-
Added the change-point information criterion
-
More options to customize the graphs from the
plot()
-method for R-objects of classhystar_fit
andhystar_sim
:-
the colors of the background and lines
-
line types and widths of
z
,y
and the thresholds
-
-
More default arguments in
z_sim
:n_t = 100
andn_switches = 2
. Now, all arguments have a default, which allows you to create examples faster. -
More default arguments in
hystar_sim
:r = c(-0.5, 0.5)
,d = 0
,phi_R0 = c(0, 0.5)
andphi_R1 = c(2, 0.5)
. Now, the only necessary argument isz
.
- Fixed a mistake in the estimation of the asymptotic standard error. I noticed this during simulations, where some confidence interval coverage rates were strikingly low. All is fine now!
- First CRAN release!