Метод к-ближайших соседей

import numpy as np
import matplotlib.pyplot as plt
 
from sklearn.metrics import euclidean_distances
from sklearn.neighbors import KNeighborsClassifier
 
from .datasets import make_forge
from .plot_helpers import discrete_scatter
 
 
def plot_knn_classification(n_neighbors=1):
    X, y = make_forge()
 
    X_test = np.array([[8.2, 3.66214339], [9.9, 3.2], [11.2, .5]])
    dist = euclidean_distances(X, X_test)
    closest = np.argsort(dist, axis=0)
 
    for x, neighbors in zip(X_test, closest.T):
        for neighbor in neighbors[:n_neighbors]:
            plt.arrow(x[0], x[1], X[neighbor, 0] - x[0],
                      X[neighbor, 1] - x[1], head_width=0, fc='k', ec='k')
 
    clf = KNeighborsClassifier(n_neighbors=n_neighbors).fit(X, y)
    test_points = discrete_scatter(X_test[:, 0], X_test[:, 1], clf.predict(X_test), markers="*")
    training_points = discrete_scatter(X[:, 0], X[:, 1], y)
    plt.legend(training_points + test_points, ["training class 0", "training class 1",
                                               "test pred 0", "test pred 1"])