R: bbmle News (original) (raw)

Changes in version 1.0.25.1

CRAN COMPATIBILITY

• remove/ignore spurious .Rout files

Changes in version 1.0.25

USER-VISIBLE CHANGES

• mle2.options() function is now exported

CRAN COMPATIBILITY

• vignettes will build without suggested packages

Changes in version 1.0.24 (2021-08-06)

USER-VISIBLE CHANGES

• new namedrop_args argument to mle2 (don't drop argument names within sub-lists of data)
• allow user-specified covariance matrix for importance sampling

Changes in version 1.0.23 (2020-02-03)

• suppress some warnings in tests for cross-platform/CRAN compatibility

Changes in version 1.0.22 (2019-12-19)

BUG FIXES

• intercept-only parameters were handled wrong (GH #8)
• fix error-handling (class vs. inherits) for R 4.0 compatibility

NEW FEATURES

• pop_pred_samp to compute population prediction samples, tweaking non-positive-definite covariance matrices and using importance samples if necessary/specified. This is still VERY experimental/in flux!
• added dnorm_n function for Normal densities with the standard deviation parameter profiled out
• add slnorm function for log-Normal densities (median and mean only for now)
• IC tables now have an as.data.frame method (from @iago-pssjd on GitHub)

USER-VISIBLE CHANGES

• the evaluatedstart and fixed components of the call are now stored in the call component (Daniel B. Stouffer)

Changes in version 1.0.20 (2017-10-30)

BUG FIXES

• fixed buglet: flipped profile plot axes, confint for negative values

USER-VISIBLE CHANGES

• summary.mle2 is now exported, for use by other packages
• small fixes to AIC(c) methods

Changes in version 1.0.19 (2017-04-08)

• fixed bug: evaluate call$method so that profiling/updating works within a function environment
• make AICtab smarter about NA values
• fix BIC bug (infinite recursion)
• hessian computation uses gradient function if provided
• basic continuation method implemented for profiling (with stubs for smarter methods)
• mle2 stores its calling environment for more flexibility when re-evaluating, e.g. in profiling (could lead to occasional surprises, e.g. if saving a fitted mle2 object with large objects in its calling environment)

Changes in version 1.0.18 (2016-02-11)

• update slice functionality; allow for explicit ranges
• CRAN updates (import from base packages)

Changes in version 1.0.17 (2014-01-01)

• new warning if ~dnorm is called with sd implicitly ==1
• some internal tweaking to slice functionality
• updated to allow for MuMIn v 1.10.0 changes

Changes in version 1.0.16 (2014-01-01)

• fix mnames behaviour
• slight vignette cleanup

Changes in version 1.0.15 (2013-11-20)

• add logLik option to IC tables, more use of nobs methods
• minor improvements to slice functionality

Changes in version 1.0.14 (2013-08-24)

• more CRAN tweaks

Changes in version 1.0.13 (2013-08-22)

• .Rbuildignore tweaks for CRAN

Changes in version 1.0.12 (2013-08-22)

• vignette tweaks
• add Depends: R>=3.0.0

Changes in version 1.0.11 (2013-08-19)

• add .Rbuildignore for CRAN tests

Changes in version 1.0.10 (2013-08-18)

• adapt to updated optimx
• tweaks for CRAN compliance

Changes in version 1.0.9 (2013-06-23)

• switch from aod to aods3 in vignettes
• switch vignette to knitr

Changes in version 1.0.8 (2013-04-23)

• tweaks to print.ICtab()

Changes in version 1.0.7

• warn on convergence failure

Changes in version 1.0.6

• fixed (fairly obscure) bug in assigning data environment to gradient function; replace 'gr' in call with appropriately evaluated version

Changes in version 1.0.5.3 (2012-09-05)

• changed some cat() to message()

Changes in version 1.0.5.2 (2012-07-29)

=======

• add .Rbuildignore for CRAN tests

Changes in version 1.0.10 (2013-08-18)

• adapt to updated optimx
• tweaks for CRAN compliance

Changes in version 1.0.9 (2013-06-23)

• switch from aod to aods3 in vignettes
• switch vignette to knitr

Changes in version 1.0.8 (2013-04-23)

• tweaks to print.ICtab()

Changes in version 1.0.7

• warn on convergence failure

Changes in version 1.0.6

• fixed (fairly obscure) bug in assigning data environment to gradient function; replace 'gr' in call with appropriately evaluated version

Changes in version 1.0.5.3 (2012-09-05)

• changed some cat() to message()

Changes in version 1.0.5.2 (2012-07-29)

>>>>>>> .r116

• remove BIC definitions (now unnecessary/should be adequately defined in core R)
• add explicit Depends: on stats4
• note that development optimx (2012.05.24+) now fails on profiling when reduced model is 1-dimensional (i.e. for a 2-parameter model)

Changes in version 1.0.5.1 (2012-07-17)

• remove spurious .RData file; version bump

Changes in version 1.0.5 (2012-05-15)

• wrapped eigen() call to prevent failure of eigenvalue ratio summary when hessian is bad
• fix bug: forgot to export summary method for mle2
• add exclude.fixed argument to coef method
• fix bug: single lower/upper parameter, or prof.lower/prof.upper parameter, not interpreted properly in profile/confint
• add and document slice methods: change from old (profile-like) behavior, old version is available as sliceOld
• DESCRIPTION/NAMESPACE fixes, move most Depends: to Imports: instead (except methods package)

Changes in version 1.0.4.2 (2012-02-25)

• fix bug in gradient/vecpar/profiling interaction (Daniel Kaschek)
• improve (and document) uniroot method for confint: now respects box constraints
• fix issue where bobyqa (optimx) strips parameter names from result

Changes in version 1.0.4.1 (2012-01-27)

• remove a bit of installed junk; vignette fix

Changes in version 1.0.4 (2012-01-02)

• fix imports for latest version of R-devel

Changes in version 1.0.3

• modified starting value code and documentation slightly: now allows per-parameter lists. Updated docs.
• Fixed bug that would screw things up if 'i' were used as a parameter.
• Update vignette to conform to MuMIn changes

Changes in version 1.0.2 (2011-09-07)

• fixed buglet in detection of no-intercept models when setting starting values

Changes in version 1.0.1 (2011-08-04)

• Turned off some warnings when skip.hessian=TRUE (e.g. in profile)
• Calculate max grad and Hessian eigenval ratio, add to "details"
• (Should add accessor methods)
• Fixed NAMESPACE to export methods etc etc properly; removed universal exportPattern directive

Changes in version 1.0.0 (2011-06-17)

• mainly just a version bump for CRAN
• added tag for variables in profile/confint warnings
• lots more technical information in vignette about profiling algorithm

Changes in version 0.9.9 (2011-05-14)

• changed NEWS to NEWS.Rd
• fixed bug for "optimize" – profile etc. now respect bounds
• eliminated warning about bounds names if all identical
• add "try_harder" flag to profiling (ignore flat spots, NAs ...)

Changes in version 0.9.8

• gradient functions work better with fixed parameters, hence with profiling
• profile plot reverts to linear for non-monotonic profile
• added warning in confint for non-monotonic profile, or for non-monotonic spline fit to monotonic profile; revert from spline+linear to linear approximation in this case
• various documentation improvements
• optimx improvements
• require data= argument when using formula interface
• turn off hessian computation in profile
• allow use of MASS::ginv

Changes in version 0.9.7

• bug fix in calc_mle2_function for no-intercept models (thanks to Colin Kremer)
• fixed optimx, added 'user' option

Changes in version 0.9.6

• changed hessian calculation to use numDeriv code (causes tiny changes to hessian results that could matter in edge cases). Too lazy to provide a backward compatibility mode ...
• documented optimizer= choices in ?mle2

Changes in version 0.9.5.1

• fixed bug in AICc (David Harris)

Changes in version 0.9.5

• added NAMESPACE, various fixes to go with that
• beginnings of an RUnit testing framework
• tweaked vignette
• added prof.lower, prof.upper to profile()
• added "optimize" to list of allowed optimizers, some bug fixes

Changes in version 0.9.4.1

• tweaked par() resetting in profile plots

Changes in version 0.9.4

• more qAICc fixing

Changes in version 0.9.3 (2009-09-18)

• tweaked handling of bounds: profile now succeeds on some 1D problems where it didn't before
• added deviance, residuals methods
• added newparams argument to predict, simulate; newdata argument to simulate
• added vignette (stub)
• added explicit params argument, to help sort out full parameter specifications when parameters is non-NULL

Changes in version 0.9.2 (2009-08-10)

• fixed predict() for case with parameters
• added snorm
• changed ICtab defaults to weight=TRUE, base=FALSE, sort=TRUE

Changes in version 0.9.1

• added simulate method (formula interface only)
• fix AICctab bug
• remove spurious cat/print in profile
• fix qAIC bug

Changes in version 0.9.0 (2008-08-26)

• fix Tom Hobbs bug: named lower/upper/parscale/ndeps get rearranged properly, otherwise rearrange in order of "start" and issue a warning
• documentation tweak for S4 as.data.frame
• added sbeta to list of known distributions
• removed nlme requirement & auto-loading

Changes in version 0.8.9 (2008-08-04)

• version bump, submit to CRAN
• added predict method

Changes in version 0.8.8 (2008-07-10)

• added flexibility for profile plotting (main, x labels etc.); added examples
• added an instance of "namedrop" to fix naming problem
• added tol.newmin to slice etc.
• added check for numeric return from profile within confint
• fixed bugs in profile plotting when profile is restricted to a subset of variables
• added tests for par() to reset to original on exit
• improved profile documentation
• replicate std.err if specified in profile
• add as.data.frame
• tweak tol.newmin (better fit found during profile) code

Changes in version 0.8.7 (2008-05-12)

• version bump, moved to R-forge.
• reordered NEWS file (most recent first)

Changes in version 0.8.6.1 (2008-03-22)

• tweaked stop-on-better-fit code
• fixed (?) qAIC(c) methods

Changes in version 0.8.6 (2008-03-26)

• tweak/fix to ICtab documentation (thanks to Tom Hobbs)
• added qAIC(c) methods (not working yet!)

Changes in version 0.8.5.1

• oops. Fixed infelicity (bug?) in new environment manipulation

Changes in version 0.8.5

• tweaked environment/data assignment to preserve original minuslogl environment better

Changes in version 0.8.4

• changed plot.profile.mle2 options (added onepage etc., made plot.confstr=TRUE by default)

Changes in version 0.8.3

• added warning about too-short lower/upper
• added documentation

Changes in version 0.8.2

• fixed bug in AICctab
• cosmetic change to printing – save call.orig
• moved ChangeLog to NEWS

Changes in version 0.8.1

fixed (?) environment bug

• tried to use built-in relist, but failed: renamed relist to "relist2" (try again later)
• documented get.mnames (auxiliary function for ICtabs)
• started to add gr (gradient) capability – NOT TESTED

Changes in version 0.8

• changed ICtab to allow either ICtab(x,y,z) or ICtab(list(x,y,z)) (L <- list(...); if is.list(L[[1]]) && length(L)==1)

Changes in version 0.7.7

• fix bug in profiling: all optim() methods EXCEPT L-BFGS-B. return the value of the objective function if given a function with no arguments/zero-length starting parameter vector (this is the situation with "profiling" a 1-D function). L-BFGS-B gives funky answers. added a check for this case. (may need to check behavior for alternate optimizers (nlm etc)) [this behavior triggered a "found better fit" error when profiling 1D functions with L-BFGS-B]
• changed behavior when finding better fit during profiling to return new parameters

Changes in version 0.7.6

• tweak vignette
• fixed second major AICc bug (was fixed in mle2 method, but not in logLik method)

Changes in version 0.7.5

• change "ll" to "LL" in examples for clarity
• tweaked anova reporting of models (wrap instead of truncating)
• added (undocumented) show.points option to profile plot to display actual locations of profile evaluation
• tweaked profile to behave better when profiling variables with constraints (upper, lower)
• moved vignette to inst/doc where it belongs
• ICtab hack to protect against package:aod definition of AIC(logLik)
• added submit stub
• tweaked slice.mle2-class docs for consistency
• fiddled with vignette
• preliminary code to allow non-monotonic profiles
• preliminary add nlm to list of optimizers (untested)
• add aod, Hmisc, emdbook to VignetteDepends and Suggests:

Changes in version 0.7

• better df extraction in ICtab
• minor bug fix for AICc (allows AICc of nls objects)
• handle models with -1 in formula better: starting values set "all equal"
• made ANOVA formula line-length accessible
• added skip.hessian and trace arguments to mle2
• messed around with BIC definition – attempt at consistency with nlme
• added rudimentary support for nlminb, constrOptim
• nlme now required for fdHess (which is required for nlminb since it doesn't compute a finite-diff Hessian)

Changes in version 0.6

• add experimental formula interface
• change all names from mle to mle2 to avoid confusion/conflicts
• with stats4 version of mle
• change internal structure of data evaluation
• worked on vignette
• added optimizer slot (stub)

Changes in version 0.5

• fix AICc bug! (was deviance+2*k*(k+1)/(n-k-1), not AIC+2*k*(k+1)/(n-k-1))

Changes in version 0.4

• change AIC to AICc for corrections
• add AICtab for weights, delta, sort ... options
• expose error messages occuring within profile()
• uniroot tries harder to find a valid endpoint
• truncate terms in anova.mle at 80 characters

Changes in version 0.3

• enhanced anova method, works with print.anova
• tweaked namedrop() code – ??

Changes in version 0.2

• added parnames, parnames<-
• minor fix to allow "profiles" of 1-parameter models (skip fdHess call)
• minor change to print method for mle results
• tweaking "vecpar" (to allow parameter vectors in objective function)
• removed fdHess/nlme dependency