Implement Hard Voting Classifier

Problem

Implement a hard voting classifier. Given predictions from multiple classifiers, the final prediction for each sample is the class that receives the most votes (majority voting).

Solution

import numpy as np
from collections import Counter

def hard_voting(predictions: list[list]) -> list:
    """
    predictions: list of arrays, one per classifier.
                 Each array has one prediction per sample.
    Returns: list of final predictions using majority vote.
    """
    n_samples = len(predictions[0])
    final_predictions = []

    for i in range(n_samples):
        votes = [clf_preds[i] for clf_preds in predictions]
        counter = Counter(votes)
        winner = counter.most_common(1)[0][0]
        final_predictions.append(winner)

    return final_predictions

Explanation

1. For each sample, collect predictions from all classifiers.
2. Count the votes for each class using a Counter.
3. The class with the most votes wins. In case of a tie, most_common returns the first one encountered.
4. Hard voting is simple and works well when individual classifiers are diverse and reasonably accurate. It does not use prediction probabilities.

Complexity

• Time: O(n * k) where n is the number of samples and k is the number of classifiers
• Space: O(k) for vote counting per sample