Add KNN classifier implementation

machine_learning/knn_classifier.py

@@ -0,0 +1,50 @@
+from collections import Counter
+from math import sqrt
+
+
+def euclidean_distance(point1: list[float], point2: list[float]) -> float:
+    """
+    Calculate the Euclidean distance between two points.
+
+    >>> euclidean_distance([1, 2], [4, 6])
+    5.0
+    """
+    if len(point1) != len(point2):
+        raise ValueError("Points must have the same dimensions")
+
+    return sqrt(sum((a - b) ** 2 for a, b in zip(point1, point2)))
+
+
+def knn_classifier(
+    training_data: list[list[float]],
+    training_labels: list[int],
+    test_point: list[float],
+    k: int = 3,
+) -> int:
+    """
+    Classify a test point using the K-Nearest Neighbors algorithm.
+
+    >>> training_data = [[1, 2], [2, 3], [3, 3], [6, 7]]
+    >>> training_labels = [0, 0, 0, 1]
+    >>> knn_classifier(training_data, training_labels, [2, 2])
+    0
+    """
+    if len(training_data) != len(training_labels):
+        raise ValueError("Training data and labels must have the same length")
+
+    if k <= 0:
+        raise ValueError("k must be greater than 0")
+
+    distances = []
+
+    for data_point, label in zip(training_data, training_labels):
+        distance = euclidean_distance(data_point, test_point)
+        distances.append((distance, label))
+
+    distances.sort(key=lambda item: item[0])
+
+    nearest_neighbors = distances[:k]
+
+    labels = [label for _, label in nearest_neighbors]
+
+    return Counter(labels).most_common(1)[0][0]