Estimates the relative overfitting of a model as the ratio of the difference in test and train performance to the difference of test performance in the no-information case and train performance. In the no-information case the features carry no information with respect to the prediction. This is simulated by permuting features and predictions.
estimateRelativeOverfitting( predish, measures, task, learner = NULL, pred.train = NULL, iter = 1 )
predish | (ResampleDesc | ResamplePrediction | Prediction) |
---|---|
measures | (Measure | list of Measure) |
task | (Task) |
learner | (Learner | |
pred.train | (Prediction) |
iter | (integer) |
(data.frame). Relative overfitting estimate(s), named by measure(s), for each resampling iteration.
Currently only support for classification and regression tasks is implemented.
Bradley Efron and Robert Tibshirani; Improvements on Cross-Validation: The .632+ Bootstrap Method, Journal of the American Statistical Association, Vol. 92, No. 438. (Jun., 1997), pp. 548-560.
Other performance:
ConfusionMatrix
,
calculateConfusionMatrix()
,
calculateROCMeasures()
,
makeCostMeasure()
,
makeCustomResampledMeasure()
,
makeMeasure()
,
measures
,
performance()
,
setAggregation()
,
setMeasurePars()
task = makeClassifTask(data = iris, target = "Species") rdesc = makeResampleDesc("CV", iters = 2) estimateRelativeOverfitting(rdesc, acc, task, makeLearner("classif.knn"))#> iter relative.overfit.acc #> 1: 1 -0.02173913 #> 2: 2 0.02127660#> iter relative.overfit.acc #> 1: 1 -0.06382979 #> 2: 2 0.06000000#>#>#>#>#>#>#>estimateRelativeOverfitting(rpred, acc, task)#> iter relative.overfit.acc #> 1: 1 0.02083333 #> 2: 2 -0.02127660