dissCompare()
comparison of two groups of data represented by a pairwise dissimilarity matrix.
seqCompare()
caused an error when only one group occurs in one of the sets. Now returns NA for such sets.
internal seq.comp()
renamed as seqxcomp()
.
missing package anchors for Rd link
targets.
All TraMineRextras plot functions: improved support of
graphical parameters. (E.g., las
and cex.axis
can
now be used to control the display of tick labels.)
seqimplic.Rd: lack of description for argument ylim
.
seqpolyads()
: time reported when core > 1
was not reliable
convert.g()
: error when path=NULL
.
convert.g()
adapted to changes in ImageMagick.
dissindic()
case indicators to select observations that are typical, atypical,
illustrative of or counterexample of a statistical association between trajectories
and a covariate/typology.
seqsplot()
, argument axes
is deprecated and replaced by xaxis
.
Argument yaxis
can now also be one of "all"
or "left"
.
seqsplot()
, ylab
can be a vector of labels by group level.
seqsplot()
supports arguments of legend
among the ...
list.
plot.stslist.surv()
, plot method for objects returned by seqsurv
displays
the ylab closer from the y-axis when yaxis=FALSE
. It also recognizes the cex.lab
argument.
seqsplot()
, seqplot.tentrop()
, seqplot.tentrop.m()
, seqedplot()
, plot.dynin()
, and plot.stslist.surv()
: default value of main
is "auto"
and NULL
suppresses the titles as expected.
seplot.entrop.m()
now in namespace.
seqedplot()
argument names and behavior made conformable with those of TraMineR:::seqplot
seqplot.rf()
argument plus.one
renamed grp.meth
seqplot.rf()
which.plot = "diss.to.med"
did not work.
seqplot.rf()
issue with unwanted ":" in plot title.
sorti()
possible conflict between column names of seqdata
and arguments passed to order
.
issue with non UTF-8 commented line in TraMineRextras.Rd
Replaced permute(.)
with sample(.)
in seqCompare
to remove dependence on gtools that has become orphaned.
seqplot.rf()
gains a new argument plus.one
to control which group sizes to increase when number of sequences is not a multiple of number of groups.
seqindic.dyn()
checked whether fstat
is a function by comparing class
with a string. Now uses inherits
.
Issue with example in seqtabstocc.Rd.
Issue with non literal item name in seqsamm.Rd.
seqcta()
organizes data for CTA analysis (Studer et al., ALCR 2018).
seqsamm()
, seqsammseq()
and seqsammeha
organize and retrieve data for SAMM analysis (Studer et al., Soc. Meth. 2018).
plot.SAMM()
plot of SAMM
object returned by seqsamm
.
seqsha()
organizes data for SHA analysis (Rossignon et al., in Ritschard & Studer, 2018).
Updated description and some Rd files.
seqindic.dyn()
gains two arguments: wstep
to control the spacing between successive windows and endmiss.as.void
to control whether sequences end at first ending missing value or at the first void.
plot.dynin()
(plot method for objects produced by seqindic.dyn
) can now plot series of summary values that end with missing values.
plot.dynin()
conf
can now be logical or the confidence probability, TRUE
being equivalent to 0.95.
TSE_to_STS()
now checks that tmin > 0
.
seqplot.rf()
returns (invisibly) group memberships.
plot.dynin()
issue with check of fstat
.
seqindic.dyn()
dynamic individual indicator values.
plot.dynin()
plot method for dynin
objects produced
by seqindic.dyn
.
seqpolyads()
degree of similarities between polyadic members
(contributed by Tim Liao)
polyads
example data set.
seqplot.tentrop()
default color now set using
qualitative_hcl
from colorspace
with the 'Dark 3'
palette.
seqCompare()
BIC and Likelihood ratio test (LRT) for comparing
two sets of sequences.
seqLRT()
LRT for comparing two sets of sequences.
seqBIC()
BIC for comparing sequence data.
seqgranularity()
two new methods "first.valid"
and "last.valid"
for sequences with missing values.
seqplot.rf()
gains a new argument which.plot
.
seqgranularity()
did not return a proper stslist
object when method="mostfreq"
.
seqedplot()
: issue with color selection when only one event.
seqedplot()
: Now calls plot.survfit
with by default
conf.int=FALSE
instead of "none"
(advised by Terry Therneau).
seqprecarity(), seqprecorr(), seqprecstart()
deleted. These functions are now in TraMineR since v 2.0-13.
seqedplot()
: argument title
is deprecated, use main
instead.
seqedplot()
: when type = 'survival'
, now supports the conf.int
argument of plot.survfit
to control the display of confidence intervals.
seqedplot()
: An issue with the joint use of the xmax
and xlim
arguments when calling plot.survfit
raised a warning since Survival
v3.0.
seqprecarity()
and seqprecstart()
gain an argument with.missing
.
seqprecarity()
: there is now a print method for the generated seqprec
object that prevents the printing of all its attributes.
seqprecorr()
: new border.effect
argument. To avoid a border
effect, when method is 'FREQ'
, 'TRATEDSS'
, or 'TRATE'
, the
computed transition frequencies or probabilities p
are adjusted
as p - p/
border.effect
when any p > 1 - .1/
border.effect
.
seqtabstocc()
gains an argument with.missing
.
seqprecorr()
: method='TRATE'
now works as expected.
seqprecarity()
, seqprecorr()
, and seqprecstart()
: errors due to bad handling of missing states.
seqedplot()
called plot.survfit
with a firstx
argument
that has been suppressed (non documented in survival news);
now uses xlim
instead.
plot.stslist.surv()
: new xlab
argument.
seqprecarity()
, seqprecorr()
, and seqprecstart()
: Important
code cleaning.
seqprecarity()
, seqprecorr()
: New argument method
that
replaces tr.type
to select the type of transition weights.
Two additional methods have been implemented, namely 'RANK'
(for rank-based transition weights) and 'ONE'
for
non-weighted transitions. Also new argument stprec
to allow
specifying rank values for method='RANK'
.
seqsurv()
: levels of spell$states were not ordered conformably with the alphabet of the state sequence object, which caused plot.ststlist.surv
to use wrong colors for the survival curves when per.state=FALSE
.
seqsplot()
: color synchronization failed when one of the states selected was unobserved in one of the groups.
seqprecorr()
: identification of non-comparable states did not work as expected.
seqprecstart()
: did not work as expected when equiv class included incomparable states.
seqsplot()
now supports the ncol
argument for controlling the number of columns in the color legend.
plot.stslist.surv()
, seqplot.tentrop()
, , seqplot.tentrop.m()
, plot.seqimplic()
, print.seqimplic()
: new argument tick.last
that when set as TRUE
allows to enforce a tick mark at the last position on the time x-axis. Has no effect when the last position is 1 + a multiple of xtstep
. Default is to use the tick.last
attribute of the state sequence object.
seqrep.grp()
: an error is raised when one of the groups has less than two cases.
Function seqgen.missing()
the number of left and right
elements to be turned into missings were inverted.
Function seqeordplot()
suppressed. It is superseded by the
function seqpcplot()
of the TraMineR package.
seqprecarity()
: computes the sequence precarity index from
a provided partial state order. The index is a
corrected complexity index. See documentation page.
seqprecorr()
: determines the correction factor used by
seqprecarity()
.
seqprecstart()
: determines the starting precarity level
of the sequence from the partial state order.
seqsurv()
: fits survival curves for states in state sequences.
There is a plot method for objects returned by seqsurv()
.
seqsplot()
: specific version of the TraMineR seqplot()
function
to plotting survival curves of states in sequences.
seqrep.grp())
renamed argument names for consistency with TraMineR.
seqrep.grp())
: when diss=NULL
, the dissimilarity matrix is now
computed by calling seqdist
with the method
and possibly related parameters specified among the ...
arguments. An error is raised when diss=NULL
and no method
is provided.
seqtabstocc())
: label synchronization issues.
Uses function TraMineR.check.depr.args()
introduced in TraMineR v 2.0-8. So, depends now on version >= 2.0-8.
STS_to_SPELL()
: Deleted. Converting to the "SPELL"
format is now a feature
of seqformat()
in TraMineR 2.0-6+.
seqrep.grp()
: Added argument with.missing
to allow computing the default distances (when diss = NULL
) for sequences with missing values.
Replaced deprecated argument names in example of seqrep.grp()
.
Now imports from RColorBrewer, survival and cluster instead of depending on those packages.
Suppressed unused import from combinat.
init.c
file.
TraMineRextras has been adapted to and requires now TraMineR 2.0-5.
TSE_to_STS()
: The two last sequences were not converted and unexpected outcome in case of a single sequence.
seqplot.rf()
, seqplot.tentrop()
, seqrep.grp()
, dissvar.grp()
, plot.seqimplic()
: adjusted argument names (main, with.legend, legend.pos, diss, ...) to conform to the TraMineR 2 naming convention.
seqrep.grp()
: Removed with.missing=TRUE
in a help example as biofam.seq
doesn't contain missing values and to conform to TraMineR 2.0-5+.
seqeformat()
, seqentrans()
, seqerulesdisc()
: Renamed 'seqe
' as 'eseq
' in internal code to conform to TraMineR 2.0-5+.
Adjusted examples in help files (sortv
, seqtabstocc
, seqrep.grp
, seqplot.tentrop
, seqplot.fr
, seqimplic
, seqgranularity
, seqgen.missing
, seqedplot
, seqentrans
) to conform to TraMineR 2.0-5+.
Function ctplot()
suppressed. It is superseded by the
function seqpcplot()
of the TraMineR package.
C++ routines are now registered.
seqimplic()
: Identifies position wise typical states
of groups of sequences when compared with all the others. There
are also plot and print methods for the outcome of the function.
seqplot.rf()
: Implements the Relative Frequency Sequence Plots
introduced in Fasang and Liao (2014)
revised help page for seqemlt()
(distance of Rousset et al. 2012).
Added required basic packages to the import statements to comply with R v3.3 requirements.
seqtabstocc()
: frequencies of state co-occurrence patterns
seqgranularity()
: new method "mostfreq"
seqedplot()
: replaced calls of TraMineR internal functions seqeage
and TraMineR.legend
with respectively
TraMineRInternalSeqeage
and TraMineRInternalLegend
(should be transparent to the end user).
Fixed issue with calling Rinternals from within extern "C" in seqstart.cpp (request of B. Ripley)
NEWS is now in Rd format into inst/NEWS.Rd
FCE_to_TSE()
: Events were not properly chronologically sorted.
TSE_to_STS()
: produced inconsistent results when events were not chronologically
sorted by cases.
First release on CRAN
seqplot.tentrop()
and seqplot.tentrop.m()
: new argument cex.legend
for controlling legend scaling factor.
seqentrans()
: new option avg.occ
; when set as TRUE
the table is
completed with average number of occurrences of the subsequences
in each sequence. The latter differs from the support for
counting methods other than COBJ. Requires TraMineR 1.8-5 or higher.
dissvar.grp()
and seqrep.grp()
now also accept distance objects as mdis
argument.
group.p
has now a weights
argument.
seqedist()
: fixed a mix between norm and interval arguments.
fixed build warning generated by calls of foreign functions.