DR-Learner Pseudo-Outcome

Problem

Compute DR-learner pseudo-outcomes: ((t - e) / (e * (1 - e))) * (y - mu_t) + mu1 - mu0, where mu_t is mu1 for treated units and mu0 for controls, and e is clipped to [1e-6, 1 - 1e-6].

Empty input returns [].

Hint

Doubly robust pseudo-outcomes stay consistent if either nuisance model is right.

Test cases

case 1

in: [[1,0],[3,2],[0.75,0.5],[2,1],[3,2]]

expected: [1,-1]

case 2

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

expected: [1,1,1]

case 3

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

expected: []

Starter codePython

def dr_learner_pseudo(treatment, outcome, propensity, mu0, mu1):
    # Your code here
    pass

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