Mann-Whitney U and Wilcoxon Rank-Sum

Problem

Return [W, U] for samples x and y, where W is the Wilcoxon rank-sum statistic (the sum of the average ranks of x within the combined sample) and U = W - n1*(n1+1)/2. Ties get average ranks. Return [0.0, 0.0] if either sample is empty.

Hint

Rank the combined data, sum the ranks belonging to x, then subtract the shift.

Test cases

case 1

in: [[1,2,3],[4,5,6]]

expected: [6,0]

case 2

in: [[4,5,6],[1,2,3]]

expected: [15,9]

case 3

in: [[1,2,2],[2,3,4]]

expected: [7,1]

case 4

in: [[5],[5]]

expected: [1.5,0.5]

case 5

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

expected: [0,0]

Starter codePython

def mann_whitney_u(x, y):
    # Your code here
    pass

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