IPW Average Treatment Effect

Problem

Return the Horvitz-Thompson inverse probability weighted ATE: (1/n) * (sum(t_i*y_i/e_i) - sum((1-t_i)*y_i/(1-e_i))). Return 0.0 for empty or mismatched input, a treated unit with e <= 0, or a control with e >= 1.

Hint

Each unit stands in for 1/e or 1/(1-e) look-alikes.

Test cases

case 1

in: [[1,1,0,0],[10,12,4,6],[0.5,0.5,0.5,0.5]]

expected: 6

case 2

in: [[1,0],[5,1],[0.5,0.5]]

expected: 4

case 3

in: [[1,0],[5,1],[0,0.5]]

expected: 0

case 4

in: [[1,0],[5,1],[0.5,1]]

expected: 0

case 5

in: [[],[],[]]

expected: 0

Starter codePython

def ipw_ate(treatment, outcome, propensity):
    # Your code here
    pass

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