Paired Wilcoxon statistical test
testPaired(e, output, rankOrder = "max", alpha = 0.05)
- Input experiment
- The output for which the tet will be performed.
- The optimization strategy, can be either maximizing "max" or minimizing "min" the target output variable.
- The significance level used for the whole testing procedure.
a testPaired object
This function performs a Wilcoxon paired test to compare the methods of an experiment consisting exactly on two of them. If more methods are present, then a multiple comparison test must be applied.
# First we create an experiment from the wekaExperiment problem and prepare # it to apply the test, we must subset it to only two methods: experiment <- expCreate(wekaExperiment, name="test", parameter="fold") experiment <- expSubset(experiment, list(method = c("J48", "NaiveBayes"))) experiment <- expSubset(experiment, list(featureSelection = c("no"))) experiment <- expReduce(experiment, "fold", mean) experiment <- expInstantiate(experiment, removeUnary=TRUE) # Then we perform a Wilcoxon test procedure test <- testPaired(experiment, "accuracy", "max") summary(test)Wilcoxon test (maximize the output named accuracy) J48 vs NaiveBayes. Test accepted: p-value: 9.3811e-02 >= 0.0500