Boxplot for the ranks distribution and control hypotheses from multiple test
- The statistical test from which the plot is generated. The functions accepts either control and pairwise multiple tests.
an experiment object
This function generates a boxplot from a testMultiple statistical test showing the ordered distrubution of rankings for each method computed for the Friedman test. If the input test features a control multiple comparison then the rejected hypotheses by the Holm methd are also indicates in the plot.
# First we create an experiment from the wekaExperiment problem and prepare # it to apply the test: experiment <- expCreate(wekaExperiment, name="test", parameter="fold") experiment <- expReduce(experiment, "fold", mean) experiment <- expSubset(experiment, list(featureSelection = "yes")) experiment <- expInstantiate(experiment, removeUnary=TRUE) # Then we perform a Friedman test included ina a testMultipleControl # test procedure test <- testMultipleControl(experiment, "accuracy") # Finally we obtain the plot plotRankDistribution(test)
 "Distribution of ranks for output \"accuracy\""cat()