Quantile Coverage Rate

Problem

Compute the empirical coverage rate of a quantile forecast: the fraction of actual values less than or equal to the matching forecast, that is the share of observations in which the quantile was not exceeded.

Return 0.0 for empty input.

Hint

A calibrated 0.9 quantile forecast should be exceeded only about one time in ten.

Test cases

case 1

in: [[10,12,14],[11,11,11]]

expected: 0.3333333333333333

case 2

in: [[5,5],[5,5]]

expected: 1

case 3

in: [[1,2,3],[3,3,3]]

expected: 1

case 4

in: [[],[]]

expected: 0

case 5

in: [[4,5,6],[4,5,5.5]]

expected: 0.6666666666666666

Starter codePython

def quantile_coverage_rate(actual, forecast):
    # Your code here
    pass

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