Causal Forest Honesty Split

Problem

Return [split_count, estimate_count] for an honest causal forest split: split_count = floor(n*split_fraction) observations choose the splits and the rest estimate the leaf effects. Return [0, 0] when n is not positive or split_fraction is outside [0, 1].

Hint

Honest trees never use the same observations for splitting and estimation.

Test cases

case 1

in: [100,0.5]

expected: [50,50]

case 2

in: [101,0.5]

expected: [50,51]

case 3

in: [10,0]

expected: [0,10]

case 4

in: [10,1]

expected: [10,0]

case 5

in: [0,0.5]

expected: [0,0]

Starter codePython

def honesty_split_counts(n, split_fraction):
    # Your code here
    pass

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