Gini impurity of a group

The Gini index measures how mixed a group of examples is: 0 when every example has the same class.

What to build

A program that computes the Gini impurity of a group from its class counts.

Requirements

  • Print gini G with 3 decimal places.

Input

One line of class counts (non-negative integers).

Output

One line.

Constraints

1 to 50 counts; their total is at least 1.

Example

Input:

5 5

Output:

gini 0.500

Why it works

Σ p² is the chance that two examples drawn at random share a class, so 1 − Σ p² is the chance they differ: 0 for a pure group, larger the more mixed it is.

Examples

Example 1 · Input
5 5
Expected output
gini 0.500
Example 2 · Input
10 0
Expected output
gini 0.000

Prompts inside input() are ignored when checking, so input("n: ") is fine.

Hints

Hint 1

Divide each count by the total to get the class shares p.

Hint 2

Gini = 1 − Σ p².

Solution

Show one correct solution (try it yourself first)
counts = list(map(int, input().split()))
total = sum(counts)
print(f"gini {1 - sum((c / total) ** 2 for c in counts):.3f}")

Lesson: Decision trees →