Hits@k

Problem

Compute the hits@k metric for link prediction.

true_edges is a list of [u, v] pairs and predictions is a ranked list of predicted [u, v] pairs. Return the fraction of true edges that appear among the first k predictions, or 0.0 when there are no true edges. Endpoint order is ignored.

Hint

Hits@k rewards ranking true edges near the top of the prediction list.

Test cases

case 1

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

expected: 0.5

case 2

in: [[[0,2],[1,3]],[[0,1],[0,2],[2,3],[1,3]],4]

expected: 1

case 3

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

expected: 0

case 4

in: [[[0,2]],[[0,1]],1]

expected: 0

Starter codePython

def hits_at_k(true_edges, predictions, k):
    # Your code here
    pass

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