NEWS.md
syntable(): Fixes in fidelity calculation when using the u-value; added Fishers’s exact test to exclude non-significant fidelitysyntable(): Added options for standardization of sample group sizehcr_resampling(): Added checks for groups limitssynsort() and syntable(): Both refer to groups instead of clusters now.hcr_resampling() to perform heterogeneity-constrained random (HCR) resampling (Lengyel, Chytrý & Tichý, 2011)syntable(): Added new options of fidelity calculation (u-value and Ochiai index); improved speed using more vectorised computationssyntable(): Added option to calculate synoptic tables from long-format vegetation data; implemented with data.table for faster performance on large databases.synsort(): Added method = "totalfreq" (order species by overall frequency) and method = "manual" (order by a user-supplied vector).synsort(): The species–sample matrix is no longer returned by default (for faster performance on large datasets); set samples = TRUE to include it.synsort(): Added cluster_order to select and order cluster columns; columns not listed are omitted.per2cov(): Fixed a mistake leading to wrong cover value assignments in specific casessyntable(): Added argument to define number of decimal places in resultsclean_matrix(): Empty cell values are transformed to 0; fixed to work correctly with non-numeric values in data tablescov2per and per2cov: Added option to transform with individual scales per sampletrans_matrix(): Added option rmchar to remove the first character of the original column namesclean_matrix() to remove species without occurrences (frequency = 0) and samples without species from a species matrix in one simple steptrans_matrix() to transpose a species matrix, while preserving correct species and sample names.merge_taxa() with increased performance
backtransform to decide whether cover-abundance values should be kept as percentage cover or back-transformed into original cover-abundance valuesdrop_zero renamed to clean_matrix and set on FALSE by defaultcov2per: Fixed an error when providing a data frame with only one column as community datasyntable: Fixed an unnecessary warning message occurring at cover value transformationx-value to presence/absence scalemerge_taxa():
drop_zero to decide whether species without occurrences or empty samples should be removed or keptsyntable() and merge_taxa() automatically repair imported tables with empty character values (““)syntable() can now handle factorial variables for defining clusters, e.g. to summarize relevés according to pre-defined categories and is more flexible regarding the format of the community matrixmerge_taxa() for merging taxa/species with identical namescov2per() and per2cov() for conversion between cover-abundance codes and percentage coverdimcheckMDS() is renamed into screeplot_NMDS() with enhanced description and progress barracurve() when freq = TRUE
specresponse()
specresponse() when showing point valuesspecresponse()
specresponse(). Full model results are returned in an (invisible) object.ordiselect()
syntable() and synsort()
ordiselect(). Now returns exact proportion of selected species. Correction in selection to axis fit limits. Variable fit now only works with factor centroids.synsort() and syntable() due to unsolved incompatibilitiesspecresponse()
racurve()
ordiselect()
syntable() and synsort()
specresponses()/specresponse() into one single function specresponse()
specresponse()
gam(), rdist())specresponse()
specresponse()/specresponses(): now works also with NMDS, includes zero values and is based on presence/absence data (logistic regression). Instead of cubic smoothing splines the function now uses GLMs/GAMs.racurve()