\itemise
in Rd
News.md
drop = FALSE
to some matrices in zeroinfl
and hurdle
functions.optim
return must equal 0 (normal convergence) in the zeroinfl
function.class(obj) == X
to inherits(obj, X)"
or is.data.frame
for compatibility with R 4.0lm.wfit
to NAMESPACE
predprob.ideal
and checkVotes
for compatibility with R 4.0dtl
files in readKH
pR2
when missing data is presentideal.R
AustralianElections
includes 2010, 2013, 2016presidentialElections
includes 2016 and 2012 Alabama resultuse.voter
option to ideal
, implementing approach from Stephen Jessee's 2016 AJPS article. See ?ideal
for details.?ideal
valgrind
in ideal's mallocsImports/Depends/Suggests
etcformat.pval
for pvalues printing in Vuong output to screenreadKH.r
presidentialElections
includes 2012postProcess
, not evaluating args in call of ideal
objectdropRollCall.Rd
hurdle
and zeroinfl
formula processing if second part of formula contained a period.COPYING
file deprecated (?), deleted from repos with r174tracex
with d>1showAll
plotting option in tracex
, change to multi
(default=FALSE
)ca2006
(thanks Arthur Aguirre)pythag
deprecated in Rmath.h
, use system hypot
instead (3/13/2011)ideal
about memory etc only come on with verbose=TRUE
(req by Stephen Jessee)partyLoyalty
(thanks to Chris Hanretty)change default prior precision for item parameters to .04 (used to be .01)
added UKHouseOfCommons
data; Example 6.9 in BASS
small bug in constrain.item
(reported by Paul Johnson)
change normalization option in ideal to generate posterior means with mean 0, sd 1
do normalization over all dimensions
typos in documentation for pseudo-R2
(thanks to Henrik P)
made gam
dependency explicit
change linear.hypothesis
to linearHypothesis
AustralianElectionPolling
itemize
-> describe
)dQuote
in Rd
files (or not)AustralianElections
with 2007 resultsiraqVote
presidentialElections
with 2008 resultsnj07
vote92
predict()
method for hurdle
object when a Poisson zero hurdle model is employedy ~ . | . - x1 - x2
were not processed correctlyfixing .Last.lib
problems
turned off MDA
option in ideal by default, seems broken, needs work
startvalues
in ideal
accept names other than xstart
(anything unique that starts with x
will do)ideal
start value argument processing; thanks Bjorn HoylandpoliticalInformation
dataMDA
option to ideal
EfronMorris
dataRockTheVote
dataCITATION
and docs
to version 1.00 accompanying JSS paperprint
and summary
methods for rollcall
and ideal
objectsideal
(discovered via odd combination of store.item
and file
)dropRollCallObject
makes prettier output, works a little more sanelys109
is now updated to its final status, example points to voteview.coms110
in examples for readKH
dropInfo
to rollcall
objects produced by dropRollCall
per request of Keith Poole (this stores information about which legislators/votes are retained by dropRollCall
)presidentialElections
data from Jackman book manuscriptigammaHDR
fixed, some typos tooy ~ x
now means y ~ x | x
and not y ~ x | 1
anymore. Thus, the same default formula processing is applied to zero-inflated and hurdle models.normalizeIdeal
found by Jeff Lewisfitted()
and predict()
methods for hurdle
and zeroinfl
for the case when an offset is specified sandwich
(2.1-0) provides estfun()
and bread()
methods, thus sandwich()
and vcovOPG()
covariances are availablecountreg
vignettepredict()
methodsextractAIC()
methodss109
predict.ideal
(for the case where the researcher did not supply burn-in; 05/15/2007).simpi
)logLik.polr
hitmiss
methodoffset
and weights
arguments have been added to zeroinfl()
and hurdle()
hurdletest()
is a new wrapper for linear.hypothesis()
from the car
package to test for the presence of hurdles.zeroinfl()
and hurdle()
).ideal
re identification and the normalize optionodTest
(wrong p-value); fixed.predict.ideal
interactive()
in ideal.r
predict.ideal
as per request of Boris Shorpredict.ideal
; after version 0.73 change to negative intercept, need change in predict.ideal.r
to matchzeroinfl()
: binomial(link = link)
is deprecated, use binomial(link = linkstr)
insteadmeanzero
option in ideal, replace with normalize
ideal
(eigen
, random
, or user-supplied list); this fixes a bug reported by Boris Shor boris@bshor.com when working with large rollcall objects"\,"
in readKH
reported by Kurt Hornick Kurt.Hornik@wu-wien.ac.atseatsVote
class and ca2006
dataideal
ideal
verbose
option to ideal
ideal.c
postProcess
gets extra referencessummary.ideal
tracex
dropRollCall
(was dropping legis.data
attributes)partyLoyalty
, called by summary.rollcall(...,verbose=TRUE)
updatex.c
, updatey.c
and xreg.c
bioChemists
data found by Bettina Guen gruen@ci.tuwien.ac.at, variable kids5
was off by 1 unit, now runs from min of zero (no kids).hurdle()
and zeroinfl()
new formula interface of type y ~ x | z
where y ~ x
specifies the count model and z
the inflation/hurdle regressors.
re-structured returned value, is now more similar to "glm" objects
extended/enhanced extractor functions
plot.ideal.1d
: better left plot margin, based on max length of legis.name
plot.ideal.2d
: inconsistent testing of presence of beta in ideal object when overlaying cutting planes
plot.ideal.Rd
: more examples (but in \dontrun
)
tracex
: bug for 2d ideal
objects
tracex
: 2d, make legend lines heavier for showAll
tracex
: for R >= 2.4
, change par()
to par(no.readonly=TRUE)
fixed typo in plot.ideal.Rd
summary.ideal
found by Keith Poole (8/8/06)ideal
section on Identification changed to reflect presence of postProcess
functionsummary.rollcall
into NAMESPACE
per request of Jeff Lewisnsl
function on Windows, changed readKH
to simply call readLines
with a try-error
wrapper per suggestion of XP-user Jeff LewisdropList
improved, per suggestion of Jeff Lewissysdata.rda
; now utility data sets such as stateinfo
and partyinfo
have to loaded via data()
in functions inside readKH
subfunctions; thanks Jeff Lewis.jitter
to plot.predict.ideal
start
to burnin
in many function that handle ideal
objectscheckStart
to checkBurnIn
in idealHelper.r
, returns keep
(logical vector)plot1d
plot1d
and plot2d
from export in NAMESPACE
codetools()
, cleaned up some redundancies, overwriting of function args inside of functions etcdtnorm
regarding return valueIDEAL
regarding double printed as introllcall
documentationdtl
file support to readKH
match.call()
, rather than carrying around copies of objects, many changes to numerous functionsdropList
introduced, many functions changed; see ?dropRollCall
and ?dropUnanimous
(rollcall
objects)extractRollCallObject
for ideal
objectscomputeMargins
for rollcall
objectsconvertCodes
operates on codes
component of rollcall
objecttracex
handles 2d traces gracefullyplot.predict.ideal
constrain.legis
and constrain.item
are .01 for unconstrained parameterss109
, dropped s102
as demonstration data, ships with packagevectorRepresentation
(rollcall
object)rollcall
class to have legis.data
and vote.data
data framesrollcall
objects have party loyalty scoresrollcall
objects have lopsided dataideal
saves some barebones summary statistics (posterior means)summary.ideal
computes standard deviationssummary.ideal
and print.summary.ideal
displays information on bills that fail to discriminateplot1d
(plot method for 1d ideal
objects) has option for all legislator namesideal
objects) has option to overlay estimated cutting linesreadKH
function and supporting data objects in R/sysdata.rda
(state.info
and party
)readKH
has extensive checks and debugging of reads from web (thanks to Kurt Kornik Kurt.Hornik@wu-wien.ac.at)ideal
examples shortercoda
to list of required packagesprint.zeroinfl
(thanks to Bettina Gruen gruen@ci.tuwien.ac.at)zeroinfl
(Bettina Gruen gruen@ci.tuwien.ac.at)sc9497
(Supreme Court sample data)summary.zeroinfl
odTest
ntable
from Jim FearonbetaHPD
Any scripts or data that you put into this service are public.
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.