K-Fold交叉验证中准确率、精确率、召回率和F1分数的正确计算方法

import numpy as np
from sklearn.model_selection import train_test_split
from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score
from sklearn.neural_network import MLPClassifier # 假设使用MLP模型

# 示例数据
X = np.random.rand(100, 10)
y = np.random.randint(0, 2, 100)
clf = MLPClassifier(random_state=42, max_iter=100) # 示例分类器
n_folds = 5

# 手动实现K-Fold（存在潜在问题）
total_accuracy = 0
total_precision = 0
total_recall = 0
total_f1 = 0

print("--- 手动K-Fold评估（不推荐） ---")
for fold in range(n_folds):
    # 每次循环都随机划分，不保证是标准K-Fold
    X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=1/n_folds, random_state=fold)

    clf.fit(X_train, y_train)
    y_pred = clf.predict(X_test)

    total_accuracy += accuracy_score(y_test, y_pred)
    total_precision += precision_score(y_test, y_pred, zero_division=0)
    total_recall += recall_score(y_test, y_pred, zero_division=0)
    total_f1 += f1_score(y_test, y_pred, zero_division=0)

print(f"平均准确率: {total_accuracy / n_folds:.2f}")
print(f"平均精确率: {total_precision / n_folds:.2f}")
print(f"平均召回率: {total_recall / n_folds:.2f}")
print(f"平均F1分数: {total_f1 / n_folds:.2f}")

import numpy as np
from sklearn.model_selection import cross_val_score
from sklearn.neural_network import MLPClassifier # 假设使用MLP模型

# 示例数据
X = np.random.rand(100, 10)
y = np.random.randint(0, 2, 100) # 假设二分类问题
clf = MLPClassifier(random_state=42, max_iter=100) # 示例分类器
n_folds = 5 # K-Fold的K值

print("\n--- 使用 cross_val_score 进行标准化评估 ---")

# 计算平均准确率
accuracy_scores = cross_val_score(clf, X, y, cv=n_folds, scoring='accuracy')
print(f'平均准确率: {accuracy_scores.mean():.2f} (标准差: {accuracy_scores.std():.2f})')

# 计算平均精确率
# 注意：对于二分类，默认是针对正类（标签为1）计算。
# 如果是多分类或需要指定正类，可能需要使用 make_scorer 或指定 average 参数
precision_scores = cross_val_score(clf, X, y, cv=n_folds, scoring='precision', error_score='raise')
print(f'平均精确率: {precision_scores.mean():.2f} (标准差: {precision_scores.std():.2f})')

# 计算平均召回率
recall_scores = cross_val_score(clf, X, y, cv=n_folds, scoring='recall', error_score='raise')
print(f'平均召回率: {recall_scores.mean():.2f} (标准差: {recall_scores.std():.2f})')

# 计算平均F1分数
f1_scores = cross_val_score(clf, X, y, cv=n_folds, scoring='f1', error_score='raise')
print(f'平均F1分数: {f1_scores.mean():.2f} (标准差: {f1_scores.std():.2f})')

# 提示：如果需要计算多分类的加权/宏平均/微平均指标，
# 可以使用 'precision_weighted', 'recall_macro', 'f1_micro' 等 scoring 字符串。
# 例如：
# f1_macro_scores = cross_val_score(clf, X, y, cv=n_folds, scoring='f1_macro')
# print(f'平均F1宏平均: {f1_macro_scores.mean():.2f}')