1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
|
context("tuneThreshold")
test_that("tuneThreshold", {
# binary classification
res = makeResampleDesc("Holdout")
lrn = makeLearner("classif.rpart", predict.type = "prob")
rf = resample(lrn, task = binaryclass.task, resampling = res, measures = list(mmce))
th = tuneThreshold(rf$pred, measure = mmce)
expect_equal(length(th$th), 1L) # threshold for positive class
expect_equal(length(th$perf), 1L) # 1d-performance value
# multiclass classification
rf2 = resample(lrn, task = multiclass.task, resampling = res, measures = list(mmce))
th2 = tuneThreshold(rf2$pred, measure = mmce, control = list(max.call = 5))
# a measure that has to be maximized
rf3 = resample(lrn, task = multiclass.task, resampling = res, measures = list(acc))
th3 = tuneThreshold(rf2$pred, measure = acc, control = list(max.call = 5))
expect_equal(th3$th, th2$th, tolreance = 0.0001)
expect_equal(length(th2$perf), 1L) # 1d-performance value
expect_equal(length(th2$th), length(getTaskClassLevels(multiclass.task))) # no. of threshold = no. of classes
expect_equal(names(th2$th), getTaskClassLevels(multiclass.task)) # threshold names = class names
expect_equal(sum(th2$th), 1L) # sum of thresholds = 1L
})
|
