from sklearn.model_selection import train_test_split<\/p>\nX_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)<\/p>\n
<\/code><\/pre>\n<\/p>\n
\u4e8c\u3001\u8ba1\u7b97\u8ddd\u79bb<\/h3>\n<\/p>\n
\u5728KNN\u7b97\u6cd5\u4e2d\uff0c\u8ba1\u7b97\u8ddd\u79bb\u662f\u6838\u5fc3\u6b65\u9aa4\u3002\u5e38\u7528\u7684\u8ddd\u79bb\u5ea6\u91cf\u65b9\u5f0f\u5305\u62ec\u6b27\u51e0\u91cc\u5f97\u8ddd\u79bb\u3001\u66fc\u54c8\u987f\u8ddd\u79bb\u548c\u95f5\u53ef\u592b\u65af\u57fa\u8ddd\u79bb\u3002<\/p>\n<\/p>\n
1\u3001\u6b27\u51e0\u91cc\u5f97\u8ddd\u79bb<\/h4>\n<\/p>\n
\u6b27\u51e0\u91cc\u5f97\u8ddd\u79bb\u662f\u6700\u5e38\u7528\u7684\u4e00\u79cd\u8ddd\u79bb\u5ea6\u91cf\u65b9\u5f0f\u3002\u5b83\u7684\u8ba1\u7b97\u516c\u5f0f\u4e3a\uff1a<\/p>\n<\/p>\n
[d(p, q) = \\sqrt{\\sum_{i=1}^{n} (p_i – q_i)^2}]<\/p>\n<\/p>\n
\u5176\u4e2d\uff0c(p) \u548c (q) \u662f\u4e24\u4e2a\u6570\u636e\u70b9\uff0c(n) \u662f\u6570\u636e\u70b9\u7684\u7ef4\u5ea6\u3002<\/p>\n<\/p>\n
import numpy as np<\/p>\ndef euclidean_distance(x1, x2):<\/p>\n
return np.sqrt(np.sum((x1 - x2)2))<\/p>\n
<\/code><\/pre>\n<\/p>\n
2\u3001\u66fc\u54c8\u987f\u8ddd\u79bb<\/h4>\n<\/p>\n
\u66fc\u54c8\u987f\u8ddd\u79bb\u662f\u53e6\u4e00\u79cd\u5e38\u7528\u7684\u8ddd\u79bb\u5ea6\u91cf\u65b9\u5f0f\u3002\u5b83\u7684\u8ba1\u7b97\u516c\u5f0f\u4e3a\uff1a<\/p>\n<\/p>\n
[d(p, q) = \\sum_{i=1}^{n} |p_i – q_i|]<\/p>\n<\/p>\n
def manhattan_distance(x1, x2):<\/p>\n return np.sum(np.abs(x1 - x2))<\/p>\n
<\/code><\/pre>\n<\/p>\n
3\u3001\u95f5\u53ef\u592b\u65af\u57fa\u8ddd\u79bb<\/h4>\n<\/p>\n
\u95f5\u53ef\u592b\u65af\u57fa\u8ddd\u79bb\u662f\u6b27\u51e0\u91cc\u5f97\u8ddd\u79bb\u548c\u66fc\u54c8\u987f\u8ddd\u79bb\u7684\u6cdb\u5316\u5f62\u5f0f\u3002\u5b83\u7684\u8ba1\u7b97\u516c\u5f0f\u4e3a\uff1a<\/p>\n<\/p>\n
[d(p, q) = \\left( \\sum_{i=1}^{n} |p_i – q_i|^p \\right)^{1\/p}]<\/p>\n<\/p>\n
\u5176\u4e2d\uff0c(p) \u662f\u4e00\u4e2a\u53c2\u6570\uff0c\u5f53 (p = 2) \u65f6\uff0c\u95f5\u53ef\u592b\u65af\u57fa\u8ddd\u79bb\u5c31\u662f\u6b27\u51e0\u91cc\u5f97\u8ddd\u79bb\uff1b\u5f53 (p = 1) \u65f6\uff0c\u95f5\u53ef\u592b\u65af\u57fa\u8ddd\u79bb\u5c31\u662f\u66fc\u54c8\u987f\u8ddd\u79bb\u3002<\/p>\n<\/p>\n
def minkowski_distance(x1, x2, p):<\/p>\n return np.sum(np.abs(x1 - x2)<strong>p)<\/strong>(1\/p)<\/p>\n
<\/code><\/pre>\n<\/p>\n
\u4e09\u3001\u9009\u62e9\u6700\u8fd1\u7684k\u4e2a\u90bb\u5c45<\/h3>\n<\/p>\n
\u5728\u8ba1\u7b97\u8ddd\u79bb\u4e4b\u540e\uff0c\u9700\u8981\u9009\u62e9\u6700\u8fd1\u7684k\u4e2a\u90bb\u5c45\u3002\u53ef\u4ee5\u4f7f\u7528\u6392\u5e8f\u7684\u65b9\u6cd5\u6765\u9009\u62e9\u6700\u8fd1\u7684k\u4e2a\u90bb\u5c45\u3002<\/p>\n<\/p>\n
def get_k_nearest_neighbors(X_train, y_train, x_test, k):<\/p>\n distances = []<\/p>\n
for i in range(len(X_train)):<\/p>\n
distance = euclidean_distance(X_train[i], x_test)<\/p>\n
distances.append((distance, y_train[i]))<\/p>\n
distances.sort(key=lambda x: x[0])<\/p>\n
neighbors = distances[:k]<\/p>\n
return neighbors<\/p>\n
<\/code><\/pre>\n<\/p>\n
\u56db\u3001\u8fdb\u884c\u6295\u7968\u6216\u5e73\u5747<\/h3>\n<\/p>\n
\u5728\u9009\u62e9\u6700\u8fd1\u7684k\u4e2a\u90bb\u5c45\u4e4b\u540e\uff0c\u9700\u8981\u8fdb\u884c\u6295\u7968\u6216\u5e73\u5747\u6765\u505a\u51fa\u9884\u6d4b\u3002\u5bf9\u4e8e\u5206\u7c7b\u95ee\u9898\uff0c\u53ef\u4ee5\u4f7f\u7528\u6295\u7968\u7684\u65b9\u6cd5\u6765\u9009\u62e9\u51fa\u73b0\u6b21\u6570\u6700\u591a\u7684\u7c7b\u522b\u3002\u5bf9\u4e8e\u56de\u5f52\u95ee\u9898\uff0c\u53ef\u4ee5\u4f7f\u7528\u5e73\u5747\u7684\u65b9\u6cd5\u6765\u8ba1\u7b97\u9884\u6d4b\u503c\u3002<\/p>\n<\/p>\n
1\u3001\u5206\u7c7b\u95ee\u9898<\/h4>\n<\/p>\n
from collections import Counter<\/p>\ndef predict_classification(X_train, y_train, x_test, k):<\/p>\n
neighbors = get_k_nearest_neighbors(X_train, y_train, x_test, k)<\/p>\n
output_values = [neighbor[1] for neighbor in neighbors]<\/p>\n
prediction = Counter(output_values).most_common(1)[0][0]<\/p>\n
return prediction<\/p>\n
<\/code><\/pre>\n<\/p>\n
2\u3001\u56de\u5f52\u95ee\u9898<\/h4>\n<\/p>\n
def predict_regression(X_train, y_train, x_test, k):<\/p>\n neighbors = get_k_nearest_neighbors(X_train, y_train, x_test, k)<\/p>\n
output_values = [neighbor[1] for neighbor in neighbors]<\/p>\n
prediction = np.mean(output_values)<\/p>\n
return prediction<\/p>\n
<\/code><\/pre>\n<\/p>\n
\u4e94\u3001\u5b8c\u6574\u5b9e\u73b0<\/h3>\n<\/p>\n
\u4ee5\u4e0b\u662fKNN\u7b97\u6cd5\u7684\u5b8c\u6574\u5b9e\u73b0\uff0c\u5305\u62ec\u6570\u636e\u9884\u5904\u7406\u3001\u8ba1\u7b97\u8ddd\u79bb\u3001\u9009\u62e9\u6700\u8fd1\u7684k\u4e2a\u90bb\u5c45\u548c\u8fdb\u884c\u6295\u7968\u6216\u5e73\u5747\u3002<\/p>\n<\/p>\n
import numpy as np<\/p>\nfrom sklearn.model_selection import train_test_split<\/p>\n
from collections import Counter<\/p>\n
def euclidean_distance(x1, x2):<\/p>\n
return np.sqrt(np.sum((x1 - x2)2))<\/p>\n
def get_k_nearest_neighbors(X_train, y_train, x_test, k):<\/p>\n
distances = []<\/p>\n
for i in range(len(X_train)):<\/p>\n
distance = euclidean_distance(X_train[i], x_test)<\/p>\n
distances.append((distance, y_train[i]))<\/p>\n
distances.sort(key=lambda x: x[0])<\/p>\n
neighbors = distances[:k]<\/p>\n
return neighbors<\/p>\n
def predict_classification(X_train, y_train, x_test, k):<\/p>\n
neighbors = get_k_nearest_neighbors(X_train, y_train, x_test, k)<\/p>\n
output_values = [neighbor[1] for neighbor in neighbors]<\/p>\n
prediction = Counter(output_values).most_common(1)[0][0]<\/p>\n
return prediction<\/p>\n
def predict_regression(X_train, y_train, x_test, k):<\/p>\n
neighbors = get_k_nearest_neighbors(X_train, y_train, x_test, k)<\/p>\n
output_values = [neighbor[1] for neighbor in neighbors]<\/p>\n
prediction = np.mean(output_values)<\/p>\n
return prediction<\/p>\n
![\"python\u5982\u4f55\u5b9e\u73b0k\u90bb\u8fd1\"](\"https:\/\/cdn-kb.worktile.com\/kb\/wp-content\/uploads\/2024\/04\/25181608\/f07bb11d-43c5-4f08-b9cb-c6e76afc32c6.webp\")
<\/p>\n