- Single replicate classes now automatically removed by
plotLDA()
.
-
plotExplanatoryHeatmap()
method for theAnalysis
class now returns the plot only if the number of plots is equal to 1. -
Removed reference to the
nCores
parameter from the documentation example ofmetabolyse()
.
- Correlation analysis results now include an absolute correlation coefficient column by which the results are also arranged in descending order.
- Console output from
imputeAll()
now suppressed.
-
Temporarily added jasenfinch/missForest as a remote until stekhoven/missForest pull request #25 is resolved.
-
The limit of the number of plotted features in
plotExplanatoryHeatmap
can now be specified using thefeatureLimit
argument. -
plotExplanatoryHeatmap()
now returns NULL and returns a message when no explanatory features are found. -
Fixed the alignment of the dendrogram branches with heat map rows in
plotExplanatoryHeatmap()
. -
Fixed
ggplot2::guides()
warning inplotFeature()
andplotTIC()
. -
Fixed bug in
explanatoryFeatures()
methods forAnalysis
class and lists where the threshold was not applied. -
Corrected the text in the modelling vignette concerning the results of using unsupervised random forest for outlier detection.
-
Package version, creation date and verbose argument added to prototype of
Analysis
class. -
All generics are now defined as standard generics.
-
Added
metrics
method forAnalysis
class. -
metrics
method for lists now ignores list elements that are not of classRandomForest
.
- Changed the
RSDthresh
argument default to 50% instead of 0.5% inQCrsdFilter
generic.
-
Added a
NEWS.md
file to track changes to the package. -
pkgdown
site now available at https://jasenfinch.github.io/metabolyseR/. -
Bug reports and issues URL at https://github.com/jasenfinch/metabolyseR/issues added to package DESCRIPTION.
-
Dedicated vignettes now available for a quick start example analysis, data pre-treatment and data modelling.
-
Function examples added to all documentation pages.
-
Unit test coverage increased to > 95%.
-
Parallel processing is now implemented using the
future
package. -
plan()
from thefuture
package is re-exported. -
RandomForest
andUnivariate
classes now inherit from class theAnalysisData
class. -
Improvements to plot theme aesthetics.
-
type
argument added toplotPCA()
,plotLDA()
,plotUnsupervisedRF()
andplotSupervisedRF()
methods for theAnalysis
class. -
"pre-treated"
for specifying type argument inAnalysis
class methods now used over"preTreated"
-
Added
clsRename()
method for renaming class information columns. -
plotMeasures()
method renamed toplotMetrics()
. -
Added
plotMDS()
,plotImportance()
andplotMetrics()
methods for lists ofRandomForest
class objects. -
Added
plotExplanatoryHeatmap()
method for lists ofRandomForest
orUnivariate
class objects. -
Renamed
keepVariables()
andremoveVariables()
methods tokeepFeatures()
andremoveFeatures()
. -
Added the helper functions
preTreatmentElements()
,preTreatmentMethods()
andpreTreatParameters()
for declaring pre-treatment parameters for theAnalysisParameters
class. -
Added the helper functions
modellingMethods()
andmodellingParameters()
for declaring modelling parameters for theAnalysisParameters
class. -
Added helper function
correlationsParameters()
for declaring correlations parameters for theAnalysisParameters
class. -
Added
binaryComparisons()
method for retrieving all possible binary class comparisons from anAnalysisData
class object. -
changeParameter()
now assigns parameter values through direct assignment. -
Added
analysisResults()
method from extracting analysis elements results from theAnalysis
class. -
Added
exportParameters()
method for exporting analysis parameters to YAML file format. -
Added
dat()
andsinfo()
accessor methods for theAnalysis
class. -
Relative standard deviation (RSD) values are now specified and returned as percentages.