Skip to contents

mlr 2.4:

CRAN release: 2015-06-13

  • WrappedModel printer was slightly improved
  • ReampleResult now stores the runtime it took to resample in a slot
  • getTaskFormula / getTaskFormulaAsString have new argument ‘explicit.features’
  • getTaskData now has recodeY = “drop.levels” which drops empty factor levels
  • option fix.factors in makeLearner was renamed to fix.factors.prediction for clarity
  • showHyperPars was removed. getParamSet does exactly the same thing
  • ‘resample’ and ‘benchmark’ got the argument keep.pred, setting it to FALSE allows to discard the prediction objects to save memory
  • we had to slightly change how the mem usage is reported in tuning and feature selection See TuneControl and FeatSelControl where it is documented what is done now.
  • tuneIrace: allows to set the precision / digits within irace (using the argument ‘digits’ in makeTuneControlIrace); default is maximum precision
  • for plotting in general we try to introduce a “data layer”, so the data can be generated independently of the plotting first, into well-defined objects; these can then be plotted with mlr or custom code; the naming scheme is always generateData and plot
  • getFilterValues is deprecated in favor of generateFilterValuesData
  • plotFilterValues can now plot multiple filter methods using facetting
  • plotROCRCurves has been rewritten to use ggplot2
  • classif.ada: added “loss” hyperpar
  • add missings properties to all ctree and cforest methods: regr/classif for ctree, regr/classif/surv for cforest, and regr/classif for blackboost
  • learner xgboost was removed, because the package is not on CRAN anymore, unfortunately
  • reg.km: added param ‘iso’
  • classif.mda: added param ‘start.method’ and changed its default to ‘lvq’, added params ‘sub.df’, ‘tot.df’ and ‘criterion’
  • classif.randomForest: ‘sampsize’ can now be an int vector (instead of a scalar)
  • plotThreshVsPerf and plotLearningCurve now have param ‘facet’

new functions

  • getTaskSize
  • getNestedTuneResultsX, getNestedTuneResultsOptPathDf
  • tuneDesign
  • generateROCRCurvesData, generateFilterValuesData, generateLearningCurveData, plotLearningCurve, generateThreshVsPerfData, plotThreshVsPerf,
  • generateThreshVsPerfData accepts Prediction, ResampleResult, lists of ResampleResult, and BenchmarkResult objects.
  • experimental ggvis functions: plotROCRCurvesGGVIS, plotLearningCurveGGVIS, plotTuneMultiCritResultGGVIS, plotThreshVsPerfGGVIS, and plotFilterValuesGGVIS

new learners:

  • classif.bst
  • classif.hdrda
  • classif.nodeHarvest
  • classif.pamr
  • classif.rFerns
  • classif.sparseLDA
  • regr.bst
  • regr.frbs
  • regr.nodeHarvest
  • regr.slim

new measures:

  • brier