Evaluation Episode Count

Problem

Compute the number of evaluation episodes needed for a target confidence margin: ceil((z * std / margin)^2). A non-positive margin returns 0.

Hint

Halving the desired margin quadruples the required episode count.

case 1

in: [2,0.5,1.96]

expected: 62

case 2

in: [1,1,2]

expected: 4

case 3

in: [0,0.5,1.96]

expected: 0

case 4

in: [2,0,1.96]

expected: 0

Starter codePython

import math
def evaluation_episode_count(std, margin, z):
    # Your code here
    pass

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