dviSolve(), which now uses the
generalised likelihood ratio (GLR) for identifications in
families with more than one missing person.dviJoint(),
swapOrientation() and formatSummary(), all
used in dviSolve().plotDVI() now allows multiple AM rows, and has a new
(experimental) argument style for choosing colouring
style.dviSim() may now do multiple simulations, and gains an
argument conditional. If TRUE, references are kept
unchanged, while the missing persons are simulated conditional on
these.fire.pedFamilias::readFam().pbapply. Use pblapply() in
dviJoint().directMatch() is much faster than before, when called
from mergePM().dvi-example.fam as a system file. Hence avoid
internet download during example (gave CRAN error).dviSolve() implementing a complete
pipeline for solving a DVI case.amDrivenDVI() implementing AM-driven
analysis.setPairing() and
excludePairing() for manually fixing or disallowing certain
pairings.findUndisputed() replace argument relax
with its negation strict, and reverse the default to be
strict = FALSE.dviSim() for simulating marker data onto a
DVI dataset.findNonidentifiable().checkDVI().mergePM() gains argument verbose.consolidateDVI().findExcluded() and
exclusionMatrix().cat() instead of message()
throughout.summariseDVI() (replaced by
print.dviData()).findUndisputed() miscounted steps if
verbose = F.jointDVI() sometimes dropped victims after
findUndisputed().The dvir package is now maintained by Magnus D Vigeland.
In print.dviData(), add info on sex of
victims/missing.
Add report to output of
findExcluded().
New function plotUndisputed().
Version 3.0.0 constitutes a major rewrite of dvir,
with many new features reflecting a broader scope of the package.
Furthermore, the syntax has been greatly simplified, due to the new
dviData container class for DVI datasets.
It should be noted that these syntax changes are not backwards compatible.
Most functions of dvir now expects a
dviData object as input.
All datasets have been regenerated as dviData
objects.
Datasets from the book Mass Identifications (Kling, Egeland, Tillmar, Prieto) have been renamed with prefix KETP.
summariseDVI() is deprecated in favour of the new
print() method for dviData objects.
As of version 3.0.0, dvir requires R >= 4.1.0 and recent versions of pedtools, forrel and pedprobr.
New S3 class dviData, and constructor with syntax
dviData(pm, am, missing).
New function plotDVI() for visualising DVI
datasets.
New function relabelDVI() greatly simplifying
relabelling tasks.
New function findExcluded() for identifying victim
samples not matching any of the missing persons - and vice
versa.
New functions directMatch() and
mergePM() for analysing and merging victim samples coming
from the same individual.
New function plotSolution() helps visualising the
output of jointDVI().
New function getFamily() for extracting the family
name (or index) of missing persons.
New internal function consolidateDVI() ensures
well-formed datasets. This is called in the beginning of all major
functions.
New function compactJointRes() simplifies the output
of jointRes().
New function familias2dvir() for parsing
.fam files written by the DVI module of Familias.
sequentialDVI() has more informative output, in sync
with findUndisputed(). While these functions do almost the
same thing, the latter is generally preferred in practice.
jointDVI() gains arguments nkeep and
maxAssign. The latter triggers a gracefully exit if the
number of assignments is too large.
jointDVI() now has a progress bar (but only when
numCores = 1).