Train/Validation/Test Sizes

Problem

Compute train/validation/test sizes for a deterministic split: n_test = int(n * test_fraction), n_val = int(n * val_fraction), and n_train is the remainder.

Return [n_train, n_val, n_test].

Hint

Floor each split size independently and give the remainder to training.

Test cases

case 1

in: [100,0.1,0.2]

expected: [70,10,20]

case 2

in: [10,0.2,0.3]

expected: [5,2,3]

case 3

in: [5,0,0.4]

expected: [3,0,2]

case 4

in: [7,0.5,0.5]

expected: [1,3,3]

Starter codePython

def split_sizes(n, val_fraction, test_fraction):
    # Returns [n_train, n_val, n_test]
    # Your code here
    pass

4 test cases · run your code against them in the editor