Propensity Weight Trim

Problem

Return the propensity weights with every value above cap replaced by cap. Return a list of zeros when cap <= 0.

Hint

Trimming extreme weights limits variance at the cost of some bias.

case 1

in: [[0.5,1.5,2],1]

expected: [0.5,1,1]

case 2

in: [[1,2],5]

expected: [1,2]

case 3

in: [[],1]

expected: []

case 4

in: [[3,3],0]

expected: [0,0]

case 5

in: [[0.2,0.4],0.3]

expected: [0.2,0.3]

Starter codePython

def trim_weights(weights, cap):
    # Your code here
    pass

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