Movatterモバイル変換

SIBER 2.1.9

• Problems with rendering the permille symbol on latest OSX systemshas meant we have had to remove all reference to the unicode symbol“U+2023” or “030” within expressions. In order to create figures withthe symbol present, you will need to do so with your own code running onyour own machine with code to work with your local OS and region/localesettings.
• Re-wrote all help files to use markdown format via Roxygen2 in placeof LaTeX format.
• DOIs correctly referenced in *.Rd files using syntax.

SIBER 2.1.8

• Added a check that the data object passed tocreateSiberObject is of classdata.frame only.If this is the case, it will coerce viaas.data.frame(data.in) and issue a warning to this effect,recommending the coercion is done before passing to this function.
• Updated PACKAGE help file construction per instructions fromCRAN.

SIBER 2.1.7

• removed suggests, depends or imports “tidyverse” and instead importspecific packages as advised is best practice
• added doi to DESCRIPTION per CRAN suggestion
• added new vignette illustrating the kapow method and associated newdatasetdata/mongooseData.rda
• fixed a bug insiberKapow which will hopefully passCRAN checks re undeclared global variables. Will pick this up on nextsubmission.

SIBER 2.1.6

• added two vignettes illustrating how to compare populations andcommunities with calculation of probabilities for differences.
• replaced packagespatstat with direct importing fromnew sub-packagesspatstat.utils andspatstat.geom

SIBER 2.1.5

• Fixed bug increateSiberObject that meant z-scorescould end up out of sync with their respective grouping variable. Thanksto Edward Doherty for pointing out this odd behaviour leading to bugdiscovery 2020/11/5. Ultimately the entire package needs to be recodedfully intidyverse.
• Fixed bug inplotSiberObject that meant both x.limitsand y.limits had to be specified to invoke manual axis setting. Eitheror can now be specified on their own with the other reverting to thedefault which is the max and min values of the corresponding data axis+/- the specifiedax.pad value which defaults to 1.
• Allow users to specify a matrix of colours to define each group andeach confidence boxplot uniquely usingsiberDensityPlot(Y, clr = ...)

SIBER 2.1.4

• fixed “vignettes with duplicate titles” warning
• added new vignettePlot-posterior-ellipses that showshow to plot multiple posterior ellipse draws on top of data usingggplot2.
• fixedfitEllipse so that it now correctly inheritsparms$n.chains andparms$n.thin frominput
• added ability to save the rawjags output to*.RData file to enable convergence diagnostics for eachellipse withinfitEllipse. This option is set in the higherlevel user-specified objectparms$save.output.
• added new vignetteTest-convergence to illustrate thisnew functionality of saving the rawjags output and testingusing Gelman diagnostics.

SIBER 2.1.3

• Contains a hotfix owing to a change to ‘spatstat’ detailedbelow
• swapped packagespatstat forspatstat.utils as per instructions from their packagemaintainers 23/03/2017
• New functionssiberCentroids,specificCentroidVectors andallCentoidVectorsadded to allow pairwise comparison of the locations of two groups usingvectors. Illustrated in an accompanying vignette.

SIBER 2.1.2

• added ability to specify custom pch point orders inplotSiberObject()
• added... to plotGroupHulls

SIBER 2.1.1

• N.B. an error during uploading to CRAN meant a release was neverlodged. I have moved on to v2.1.2 for the next release as above.
• hotfix for bug in bayesianOverlap - thanks Mark Nowak for spottingthis.
• Important: install from github if you want to usebayesianOverlap for now until I can push a hotfix to CRAN.devtools::install_github("andrewljackson/SIBER", build_vingettes = TRUE)
• examples added tomaxLikOverlap() andpointsToEllipsoid().

SIBER 2.1.0

• Added functions to calculate whether arbitrary points are inside oroutside ellipses or higher dimensional ellipsoids along withillustrative vignettes.
• Changed method of calculation of angle of ellipse with x-axis tousingatan in place ofasin which is a moreelegant way of ensuring the sign of the returned angle is correct.
• Small sample size correction for drawing ellipses can now be toggledusing addEllipse(small.sample = TRUE, m = m) effectively meaning SEAc orSEA can be illustrated.
• New vignette added illustrating how to calculate overlap between twoellipses. Two new functions detailed below greatly improve the ease withwhich this can be applied.
• Fixed three bugs in the ellipse overlap vignette. Thanks toSarina forpointing this out. These are no longer an issue as the new functionsdetailed below replace this code in the vignette, but it was helpful forme during the creation of these functions. Thanks.
• Added a new functionbayesianOverlap() that calculatesthe posterior overlap between ellipses fitted to two groups. Thanks toJosh Stewart for forcing my hand on this long-overdue feature.
• Added a new functionmaxLikOverlap() to ease thecalculation of overlap between ellipses using the ML estimated ellipseswhich previously required more manually coding than was ideal. ThanksMark Nowak and SarinaJ for spotting some bugs and helping work themout.
• added new functionsellipseInOut() andpointsToEllipsoid to enable testing of whether points lieinside or outside an n-dimensional ellipsoid, including the bivariateellipse. These are useful for testing that the quantile predictionellipses do indeed contain the expected number of data points from asample. It might also be useful for assignment, identification ofoutliers, or measures of overlap of individual data points with otherellipses.

SIBER 2.0.3

• Added a new vignette illustrating how to add custom ellipses to eachgroup manually using the functionaddEllipse

SIBER 2.0.2

• Bug in Group labels as character strings fixed
• Community labels as character strings now implemented

SIBER 2.0.1

• Group labels can now be strings and do not have to be sequentialintegers (#14)

SIBER 2.0

• Major overhaul of all code and underlying fitting algorithms
• Fitting is now via JAGS
• Data are z-score transformed prior to fitting to improveconvergence
• Data structures have now changed from previous versions embeddedwithin SIAR so you will have to reformat your data and write new scriptsto interface with the new code