ROC曲线案例（了解）

import numpy as np
import matplotlib.pyplot as plt
from sklearn.metrics import roc_curve, auc

# 1. 准备数据
# 6个样本：2个正例（被点击），4个负例（未被点击）
y_true = np.array([1, 1, 0, 0, 0, 0])  # 真实标签
y_prob = np.array([0.9, 0.8, 0.7, 0.6, 0.5, 0.4])  # 预测概率

# 2. 计算ROC曲线
fpr, tpr, thresholds = roc_curve(y_true, y_prob)

# 3. 计算AUC
auc_value = auc(fpr, tpr)
print(f"AUC值: {auc_value:.4f}")

# 4. 打印每个阈值的结果
print("\n=== 各阈值结果 ===")
print("阈值\tFPR\tTPR")
for i in range(len(thresholds)):
    print(f"{thresholds[i]:.2f}\t{fpr[i]:.4f}\t{tpr[i]:.4f}")

# 5. 绘制ROC曲线
plt.figure(figsize=(10, 8))
plt.plot(fpr, tpr, color='darkorange', lw=2, 
         label=f'ROC curve (AUC = {auc_value:.4f})')
plt.plot([0, 1], [0, 1], color='navy', lw=2, linestyle='--', 
         label='Random (AUC = 0.5)')

# 标注坐标点
for i in range(len(thresholds)):
    plt.annotate(f'({fpr[i]:.2f}, {tpr[i]:.2f})', 
                (fpr[i], tpr[i]), 
                textcoords="offset points", 
                xytext=(10,10), 
                ha='center')

plt.xlim([0.0, 1.0])
plt.ylim([0.0, 1.05])
plt.xlabel('False Positive Rate (FPR)')
plt.ylabel('True Positive Rate (TPR)')
plt.title('ROC Curve - 广告点击预测')
plt.legend(loc="lower right")
plt.grid(True, alpha=0.3)
plt.show()

# 6. 手动计算验证
print("\n=== 手动计算验证 ===")
thresholds_manual = [0.9, 0.8, 0.7, 0.6, 0.5, 0.4]

for threshold in thresholds_manual:
    y_pred = (y_prob >= threshold).astype(int)
    
    # 计算混淆矩阵
    tp = np.sum((y_pred == 1) & (y_true == 1))
    fp = np.sum((y_pred == 1) & (y_true == 0))
    fn = np.sum((y_pred == 0) & (y_true == 1))
    tn = np.sum((y_pred == 0) & (y_true == 0))
    
    fpr_val = fp / (fp + tn) if (fp + tn) > 0 else 0
    tpr_val = tp / (tp + fn) if (tp + fn) > 0 else 0
    
    print(f"阈值={threshold}: TP={tp}, FP={fp}, FN={fn}, TN={tn}, "
          f"FPR={fpr_val:.4f}, TPR={tpr_val:.4f}")