Bug fixes:
nlcpd_self_start(strategy="s1")
where the base
region was not set properly.pricelevels(settings=list(solve="matrix"))
where the
setting was not supported and is now internally set to
iterative
.nlcpd(settings=list(use.jac=TRUE)
when there were
two regions (R=2
) leading the matrices B
and
C
in nlcpd_jacobi()
to collapse. Solved by
adding drop=FALSE
to the two matrices and added tests for
this case.lehr()
, when some quantities in the data were 0, the
function returned NaN
.minpack.lm
-package in the
documentation for cpd()
and nlcpd()
.Improvements and other changes:
pricelevels(settings=list(type=c("a","b"")))
, no
alphabetic ordering by index name of the output if specific indices are
provided; alphabetic ordering only if
settings$type=NULL
.geoyoung()
for the geometric Young
index. Function description, tests and vignette updated.all.pairs=FALSE
, the functions
index.pairs()
and geks()
no longer use the
price levels where the comparison and base region are the same (e.g.,
\(P^{11}\)) as these price levels do
not provide valuable interregional information.attr(x, "base")
is always added to the
function output of ratios()
if static=FALSE
and base is not NULL
. The attribute provides the base
region used to calculate the individual price ratios for each product.
The treatment of duplicated prices by ratios()
is now also
aligned to the respective treatment by all index formulas in the
package, that is, duplicated prices are averaged and ratios derived by
dividing the initial prices by the average.pairs()
to compute the number of
non-redundant index pairs available for a data set. Renamed the
functions comparisons()
and sparsity()
to
properties()
and gaps()
. Updated the output of
properties()
.settings$prods
, which removes products with prices in only
one region. This is needed because otherwise these products distort the
estimation by cpd()
, ikle()
,
gerardi()
, and others if weights are used.options(pricelevels.[])
plot.pricelevels()
ratios()
arrange()
mjevons()
in rao()
mcarli()
in rhajargsaht()
mdutot()
in gkhamis()
mharmonic()
in ikle()
q=NULL
and/or w=NULL
mjevons()
, mcarli()
,
mdutot()
, and mharmonic()
nlcpd()
, introduced matching of lower and upper
bounds and replaced argument par.start
with the default
par
gerardi()
index from the iterative
onesarrange()
In the DESCRIPTION file:
\dontrun{}
with \donttest{}
in
the examples.