changes in version 2.4.4 (2018-05-03)
*finally updated my email in package description
* set zero=FALSE in eigen.analysis to match default in sensitivity
changes in version 2.4.3 (2016-03-31)
*fixed error in demo(fillmore) and demo(stage.classify) due to old code from demo graphics
changes in version 2.4.2 (2015-11-02)
*fixed Notes required by CRAN, export mean.list and removed invalid URLs in help pages
changes in version 2.4.1 (2015-11-02)
*added secder and stoch.sens function in the demogR package since it was removed from CRAN
*fixed bug in y-axis labels in logi.hist.plot when boxp=TRUE
*changed imports to suggests quadprog package. This package is required for QPmat only
*removed invalid URLs in some help pages
changes in version 2.4 (2012-03-20)
*removed vignette due to errors with greek symbols, unloaded packages
(lattice, XML) and others. The old vignette examples are now
at https://github.com/cstubben/popbio/wiki
*replaced sd(logLam) with apply(logLam, 2, sd) in vitalsim to avoid
Warning: sd() is deprecated
* changed stoch.projection to only use a list of matrices. The matrices
are now sampled once before the loop and the code runs much faster.
* fixed bug in boot.transitions if some stages are NA. Vectors
will have an extra column and return an unhelpful error.
* fixed bug in resample if splitting A matrix using a F matrix
(i.e, some elements have both T and F values and were not added)
changes in version 2.3.1 (2011-7-11)
* changed option strip.white=FALSE to strip.white=false in
load_matrix.Rnw (SweaveOpts line) to avoid warning in R CMD check
changes in version 2.3 (2011-6-22)
* added matrix2 function to create square matrices and updated vignette
changes in version 2.2 (2011-3-22)
* added srt to control label rotation in image2
* compressed aq.trans and aq.census to avoid warnings in new R version
changes in version 2.1 (2010-9-30)
* added option to split A using a logical matrix since row and
column indexes will not always work.
* created vignette describing two ways (using scan or read.table)
to load projection matrices into R. Try vignette("load_matrix")
changes in version 2.0 (2009-9-14)
* split the six combined parts of eigen.analysis into separate
functions (lambda, stable.stage, reproductive.value,
damping.ratio, sensitivity, and elasticity). Among other things,
this makes it easier to apply the functions to a list of
matrices.
* modified the three functions that calculate age-specific traits
(generation.time, net.reproductive.rate and fundamental matrix) to
accept a projection matrix as input (which is then split into
separate T and F matrices using the splitA function). Again, this
makes it easier to apply the functions to a list of matrices.
* modified the 3 main datasets (teasel, tortoise, and whale) to
include only a projection matrix (and not separate T and F
matrices, since A can be split using splitA)
* added splitA function to decompose a projection matrix into T and
F matrices
* added resample function to resample a projection matrix
using a multinomial distribution for transitions and a log normal
distribution for fertilities
* added var2 function to calculate the variance of a list of
matrices (like mean.list, but variance is not generic, so var2)
* added matplot2 function to plot the rows of a matrix with a
legend included
* moved a number of long help examples that re-created plots in
Caswell (2001) to a single demo called Caswell.
* minor changes include updating lnorms, allowing NAs in image2
and projection.matrix, adding pooled matrix to calathea, changing
defaults for image2.
changes in version 1.1.11 (2008-12-10)
* added image2 plot for projection, sensitivity and elasticity
matrices. Examples were added to a number of help files.
* fixed stage.vector.plot to allow NA values as part of input
* forgot to escape matrix multiplier %*% in LTRE example, so last
plot and matrix were wrong, since % is intrepreted as comment
character in help files (*.Rd).
changes in version 1.1.9 (2008-07-08)
* fixed error in vitalsens example noted by Jay Rotella.
Previously, Sf3 is set to 1 in last loop and then not reset to
.8943 when calculating vitalsens(goose.el, goose.vr).
* changed x@gradient[i] to attr(x, "gradient")[i] in vitalsens to
avoid error in R-devel 2.8: trying to get slot "gradient" from an
object of a basic class ("numeric") with no slots
* added ChangeLog and Citation. Now citation("popbio") will
return JSS article.
* saved hudcorrs as *.rda file type with two elements, corrin
and corrout (ie, removed vector of vital rate labels).
* set varF=NULL as default in multiresultm
* fixed test.census and projection.matrix examples to match names
used in JSS paper (eg, trans01 was previously named test.trans and trans)
changes in version 1.1.8 (2008-03-13)
* added grizzly bear population counts for count-based PVAs
* added extCDF function for extinction time cdfs. This function
is called by countCDFxt and was previously hidden in that
function.
* added plot option to countCDFxt and set default value for tq
option, which is usually equal to number of tranistions (nt) for
regularly censused populations.
changes in version 1.1.7 (2008-02-21)
* added Calathea to LTRE example with plot, year, and interaction
effects
* added s3 method for function mean.list to get the mean matrix
from a list of matrices.
changes in version 1.1.6 (2008-01-14)
* changed ltre to LTRE
* added plots to vitalsens example
changes in version 1.1.5 (2008-01-05)
* added ltre function for fixed LTRE. A random design is included
in the example
* added Calathea dataset for tropical understory herb in Horvitz
and Schemske (1995)
* fixed damping ratio in eigen.analysis. Now the function finds
the second largest modulus and not just the modulus of the second
eigenvalue ( in case of eigenvalues with equal magnitudes in
imprimitive matrices). Also, added which.max to find dominant
eigenvalue instead of which(Re(ev$values)==max(Re(ev$values)))
which may return two or more indexes in some cases
changes in version 1.1.4 (2007-11-13)
* fixed error in image plot on teasel example after loading to
CRAN
changes in version 1.1.3 (2007-11-13)
* added function vitalsens for vital rate sensitivities and
elasticities
* added matplots to teasel example
* changed name of lambda to lambda1 in eigen.analysis output to
match package demogR (in case demogR is loaded after popbio, then
functions using eigen.analysis should still work with some
exceptions, eg, if matrix is singular)
changes in version 1.1.2 (2007-09-13)
* fixed undefined nyrs variable in varEst
* added checks to head2 function to work with matrix or dataframe
input only
* removed extra brackets, foreign language encodings (Akcakaya) in
.Rd files requested by JSS editors
changes in version 1.1.1 (2007-08-25)
* added pfister.plot to create log-log plots of variance
vs. sensitivity and CV vs. elasticity
* added more plotting examples (fig 9.3, 9.4, 9.11) from chap 9 in
Caswell to teasel and tortoise datasets
* added age-specific survival and fertility curves (fig 5.1 and
5.2) to whale dataset
* corrected errors in hudcorr matrices and vitalsim function
suggested by Doak in email dated 8/4/07.
* added plot based on fig 8.3 to vitalsim function and code (not
run) to simulate no correlations or only within-year correlations
* changed colors in stage.vector.plot to rainbow(8)
* added column names to boot.transitions (vector and matrix
output, e.g., a11, a21, a31 a12, ...) and pop.projection output
* changed stage.vector.plot examples to project matrix in example
2.1 in Caswell and create figure 2.3.
changes in version 1.1 (2007-07-28)
* added new functions and datasets contributed by Patrick Nantel
* added default row and column names (if missing) to
stage.vector.plot, checks for matrix with 2 or more stages, and
repeats colors if necessary
* now checking if matrix is singular in eigen.analysis. If matrix
is singular, NA values are used instead of exiting with a warning.
This is useful for bootstrapping that may create singular matrices
* added print statements on long loops used in all three
stochastic growth functions to track progress
* now rounding stages first in vector before checking extinction
probability in stoch.quasi.ext. Now if vector is (.3,.4,.4) and
ext prob=1 then rounding stages sets n=0.