ROC Curve Points

Problem

Compute ROC curve points for binary labels. Start at [0.0, 0.0] (threshold above every score), then for each unique score in descending order predict positive when score >= threshold and append [FPR, TPR].

If there are no positive or no negative labels, the corresponding rate is 0.0. Empty input returns [].

Hint

Sweep the unique scores from high to low and track cumulative TP and FP fractions.

Test cases

case 1

in: [[0,0,1,1],[0.1,0.4,0.35,0.8]]

expected: [[0,0],[0,0.5],[0.5,0.5],[0.5,1],[1,1]]

case 2

in: [[1,1,0,0],[0.9,0.8,0.4,0.1]]

expected: [[0,0],[0,0.5],[0,1],[0.5,1],[1,1]]

case 3

in: [[1,1],[0.5,0.5]]

expected: [[0,0],[0,1]]

case 4

in: [[],[]]

expected: []

Starter codePython

def roc_curve_points(y_true, scores):
    # Your code here
    pass

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