Combine two experiments with different outputs
expCombine(e1, e2, name = NULL)
- First experiment to combine.
- An second experiment to combine, must share the same config as e1.
- Optional name for the resulting experiment. If not specified the new experiment will be called "e1_name U e2_name"
An new experiment with common rows and all columns.
This fuctions joints two experiments sharing the same configuration of methods, problems and parameters but different outputs. The resulting experiment includes the common rows for both experiments with all the output columns.
# In this example we turn the wekaExperiment into two different experiments, # with different outputs to combine them: df_acc <- wekaExperiment[, c("method", "problem", "fold", "featureSelection", "accuracy")] df_time <- wekaExperiment[, c("method", "problem", "fold", "featureSelection", "trainingTime")] exp_acc <- expCreate(df_acc, name="acc", parameter="fold") exp_time <- expCreate(df_time, name="time", parameter="fold") # With expCombine we can mix the two experiments: expCombine(exp_acc, exp_time)#Experiment name: acc U time #method: J48, NaiveBayes, OneR, RandomForest #problem: anneal, audiology, balance-scale, car, glass, horsecolic, hypothyroid, ionosphere, liver-disorders, lymph, primary-tumor, soybean, vehicle, vote, vowel #parameters: *) fold [0,1,2,3,4,5,6,7,8,9] *) featureSelection [no,yes] #outputs: accuracy, trainingTime