R-Learner Pseudo-Outcome

Problem

Compute R-learner pseudo-outcomes: (y - m) / (t - e), clipping the treatment-minus-propensity denominator to at least 1e-6 in magnitude with its original sign.

Empty input returns [].

Hint

The pseudo-outcome regresses the residualized outcome on the residualized treatment.

Test cases

case 1

in: [[2,4,6],[1,2,3],[1,1,0],[0.5,0.5,0.5]]

expected: [2,4,-6]

case 2

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

expected: [8]

case 3

in: [[3],[1],[1],[1]]

expected: [2000000]

Starter codePython

def r_learner_pseudo(y, m, treatment, propensity):
    # Your code here
    pass

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