Skip to content

Latest commit

 

History

History
134 lines (104 loc) · 5.61 KB

NEWS.md

File metadata and controls

134 lines (104 loc) · 5.61 KB

version 6.18.0

new features / enhancements / changes

  • new function plotMarkers to visualise the selected features in block analyses (see #134)
  • auroc title now fixed (#135)
  • cimDiablo takes trim argument to customise outlier filtering (#136)
  • plotIndiv.pca default shape set to 16
  • circosPlot & network now support blocks with similar feature names
  • circosPlot now has methods for block.spls objects
  • circosPlot now takes new formal and advanced args for customisation. See ?circosPlot.

bug fixes

  • plotVar legend colour mismatch bug fixed
  • plotDiablo error undefined variable (Y) fixed
  • nipals initialisation bug with high-variance high-NA rate column fixed
  • cim bug with high NA rate data fixed using imputation by the column mean

version 6.16.0

new features / enhancements / changes

  • biplot default colours now match plotIndiv
  • biplot now takes cex argument
  • nipals now takes center and scale arguments
  • nipals now only outputs p, t and eig
  • new function impute.nipals to impute missing values using NIPALS algorithm
  • nipals function checks for orthogonality of components when NA present
  • plot.tune legends improved
  • plot.tune now uses colour-blind friendly colours
  • new tune.spls function to perform variable selection on both X and Y
  • tune.spls now chooses optimal keepX even if nrepeat < 3
  • tune.spls now takes validation arg
  • tune.spca is now much faster
  • circosPlot links can now have adjustable width (#118)
  • plotDiablo now takes col.per.group (#119)
  • terminology change: consensus renamed to average in Diablo context

bug fixes

  • plotVar bug fixed

version 6.14.0

new features / enhancements / changes

  • circosPlot: The radial location of feature names can now be cutomised using var.adj
  • added plot and print methods for nipals ouput (#87)
  • all Discriminat Analyses now run solely on mode=regression (#79)
  • cim argument change: threshold replaced by cutoff
  • nipals and pca with missing values allow skipping reconstitution of the input matrix
  • tune.block.splsda now allows random number seed also for parallel processing (#72)
  • New biplot methods for the pca family (#90)

bug fixes

  • plotIndiv: Legend bug which misspecified the groups resolved
  • plotIndiv: Legends now ordered as inputted, and not alphabetically
  • plot method issue for spca resolved
  • plotLoadings.spca bug with var.names now fixed (#81)
  • ipca deprecation warning fixed

version 6.12.0

new features / enhancements

  • plotLoadings's infamous figure margins too large error now handled and informative condition thrown
  • circosPlot's lines argument default to FALSE now
  • circosPlot's inconsistentcy of blocks with identical X names fixed
  • consensus and weighted consensus plots now supported for plotIndiv with relevant block analyses
  • plotLoadings's feature name trimming can be customised
  • block.splsda bug which could drop some Y factors with near.zero.variance=TRUE fixed
  • perf.block.splsda now supports calculation of combined and per-block AUC
  • model improvement significance can be custmoised in all perf and tune functions
  • perf.block.splsda is now much faster and supports FORK clusters
  • tune.(s)pls(da), perf.(s)plsda now support FORK clusters

bug fixes

  • circosPlot's faded lines bug when many NAs present fixed
  • tune.block.splsda() bug when using fixed test.keepX over two or more blocks fixed
  • circosPlot and plotLoadings bug caused by features with NAs fixed
  • plotIndiv(..., ind.names = FALSE) warning/bug fixed
  • tune.block.splsda bug on Windows parallelisation fixed
  • perf and tune functions' issue when choosing the optimum component resolved
  • added option to suppress auroc from printing all the AUCs

version 6.10.0

new features / enhancements

  • parallel processing on tune.block.splsda improved
  • tune.block.splsda now supports more distances
  • You can now customise auroc plots. Refer to documentation for more info

bug fixes

  • single factor multilevel error in pls fixed
  • fixed over-estimated correlation of cim for mixo_(s)pls objects with single component
  • margin error in cim now handled properly
  • fixed plotLoadings error for very long variable names
  • predict function bug for single sample prediction fixed
  • plotLoadings bug for long variable names fixed
  • Fixed tune.spls and pef.plsda bugs when using cpus argument for parallel processing
  • perf.plot bug in extracting names fixed
  • Few fixes for tune.splsda with AUC

minor improvements

  • missing values in plotIndiv's group argument no more throws error
  • mixOmics::predict function documentation now more accessible
  • names of linnerud datasets fixed
  • plot.perf now respects ylim arguments for custom y range
  • package startup message with direct liks to useful resources
  • mixOmics function documentation disambiguated with instruction on how to get package help
  • Updated onLoad message with discussion forum info, bug reports, and more
  • Dropped legacy comp.tol argument from pca
  • plot.perf now respects ylim arguments for custom y range
  • Added Code of Conduct

version 6.8.0

  • NOW HOSTED ON BIOCONDUCTOR

version < 6.8.0

  • Refer to ./inst/legacy/NEWS-old on GitHub repo