| NEWS | R Documentation |
Runtimings now added to output object.
More testing of fall back options.
sampfrac argument deprecated. Only values of 1 now accepted. This never did work properly and will be removed in future releases.
OpenBugs no longer supported.
Diagonal model now allows you to fix some P's – typically used when second trap is not running.
All plots converted to ggplot format and stored in the plots object in the final result. An argument is available in the fitting routines to stop saving output to files.
Bayesian p.value plot created for MarkAvail routines (issue #17)
Several vignettes created to replace and enhance demo files.
Corrected error in creating u2copy to improve mixing. Issue #18..
Corrected error where output of Theta in non-diagonal cases was stored incorrectly in sims.matrix and sim.array objects. Code used sim.list where Theta were stored correctly, so no effect on fitting.
Bayesian P-values plots are now created using ggplot() and saved in results.
Bayesian p-value plots were incorrect for hatchery/wild mixtures.
Updated Imports/Suggests to conform to latest CRAN standards
Now uses JAGS rather than OpenBugs so BTSPAS should run on all platforms. OpenBUGS can still be run using the 'engine' argument on the *_fit() call.
Minor changes to package file structures to meet R package conventions. No change in functionality.
Minor bug fixes discovered in conversion to JAGS.
Initialized missing values in n1, u2, m2 to sensible values to avoid OpenBugs from wandering too far off and generating nonsense values.
Corrected initialization of epsilon when logitP.fixed is set
Minor changes to package file structures to meet R package conventions. No change in functionality.
None.
Plots in non-diagonal case use all time periods rather than just those from releases.
Cleanup of some function calls; Fix arguments for xlim in some plots; don't try and print Rhat if number of chains is 1.
NONE
Fixed and issue where if you had missing values in your data, or zero in n1 and m2 (which the program automatically changes to NA), the scaling factor to compute initial values was returned as NA which caused lm() to fail when initial values are generated.
Posterior Predictive Plots modified to make axes symmetric.
Predictive posterior plots (Bayesian P-value plot) did not have symmetric X/Y axis so the p-value may not print. The axes are now forced to be symmetric.
Misc changes to correct typos, documentation, etc.
BTSPAS now uses OpenBugs exclusively so it can run on Windoze, Mac, and Unix. Users may need to install OpenBugs in place of WinBugs. No other change in functionality.
NONE.