{"id":979801,"date":"2024-12-27T06:49:24","date_gmt":"2024-12-26T22:49:24","guid":{"rendered":"https:\/\/docs.pingcode.com\/ask\/ask-ask\/979801.html"},"modified":"2024-12-27T06:49:26","modified_gmt":"2024-12-26T22:49:26","slug":"python-lift%e6%9b%b2%e7%ba%bf%e5%a6%82%e4%bd%95%e7%94%bb","status":"publish","type":"post","link":"https:\/\/docs.pingcode.com\/ask\/979801.html","title":{"rendered":"python lift\u66f2\u7ebf\u5982\u4f55\u753b"},"content":{"rendered":"

\"python<\/p>\n

Python\u4e2d\u7ed8\u5236Lift\u66f2\u7ebf\u53ef\u4ee5\u901a\u8fc7\u4f7f\u7528scikit-learn\u5e93\u3001matplotlib\u5e93\u548cpandas\u5e93\u5b9e\u73b0\u3002\u9996\u5148\u9700\u8981\u901a\u8fc7\u9884\u6d4b\u6982\u7387\u5bf9\u6837\u672c\u8fdb\u884c\u6392\u5e8f\uff0c\u7136\u540e\u8ba1\u7b97\u6bcf\u4e2a\u5206\u4f4d\u6570\u7684\u7d2f\u8ba1\u54cd\u5e94\u6bd4\u4f8b\u548c\u57fa\u51c6\u6bd4\u4f8b\uff0c\u6700\u540e\u7ed8\u5236\u51faLift\u66f2\u7ebf\u3002\u4f7f\u7528scikit-learn\u5e93\u4e2d\u7684\u51fd\u6570\u53ef\u4ee5\u7b80\u5316\u8ba1\u7b97\u8fc7\u7a0b\uff0cmatplotlib\u5e93\u5219\u7528\u4e8e\u7ed8\u5236\u56fe\u5f62\u3002<\/strong> <\/p>\n<\/p>\n

\u4e3a\u4e86\u8be6\u7ec6\u63cf\u8ff0\u5176\u4e2d\u7684\u4e00\u4e2a\u6b65\u9aa4\uff0c\u6211\u4eec\u53ef\u4ee5\u7740\u91cd\u8bb2\u89e3\u5982\u4f55\u901a\u8fc7\u9884\u6d4b\u6982\u7387\u5bf9\u6837\u672c\u8fdb\u884c\u6392\u5e8f\u3002\u9996\u5148\uff0c\u901a\u8fc7\u6a21\u578b\u7684\u9884\u6d4b\u6982\u7387\uff0c\u6211\u4eec\u53ef\u4ee5\u83b7\u5f97\u6bcf\u4e2a\u6837\u672c\u88ab\u9884\u6d4b\u4e3a\u6b63\u7c7b\u7684\u6982\u7387\u3002\u7136\u540e\uff0c\u6211\u4eec\u5c06\u8fd9\u4e9b\u6982\u7387\u4ece\u9ad8\u5230\u4f4e\u8fdb\u884c\u6392\u5e8f\uff0c\u8fd9\u6837\u505a\u7684\u76ee\u7684\u662f\u4e3a\u4e86\u8ba9\u6211\u4eec\u5728\u8ba1\u7b97Lift\u503c\u65f6\uff0c\u80fd\u591f\u4f18\u5148\u8003\u8651\u90a3\u4e9b\u9884\u6d4b\u4e3a\u6b63\u7c7b\u6982\u7387\u66f4\u9ad8\u7684\u6837\u672c\uff0c\u4ece\u800c\u66f4\u51c6\u786e\u5730\u8bc4\u4f30\u6a21\u578b\u7684\u6027\u80fd\u3002<\/p>\n<\/p>\n

\u4ee5\u4e0b\u662f\u8be6\u7ec6\u7684\u5185\u5bb9\u5206\u6bb5\uff1a<\/p>\n<\/p>\n

\u4e00\u3001LIFT\u66f2\u7ebf\u7684\u6982\u5ff5\u548c\u5e94\u7528<\/h3>\n<\/p>\n

Lift\u66f2\u7ebf\u662f\u7528\u4e8e\u8bc4\u4f30\u5206\u7c7b\u6a21\u578b\u6027\u80fd\u7684\u5de5\u5177\uff0c\u7279\u522b\u662f\u5728\u4e8c\u5206\u7c7b\u95ee\u9898\u4e2d\u3002\u5b83\u5e2e\u52a9\u6211\u4eec\u7406\u89e3\u6a21\u578b\u5728\u4e0d\u540c\u5206\u4f4d\u6570\u4e0a\u7684\u8868\u73b0\uff0c\u8861\u91cf\u6a21\u578b\u7684\u9884\u6d4b\u80fd\u529b\u3002<\/p>\n<\/p>\n

1\u3001LIFT\u66f2\u7ebf\u7684\u5b9a\u4e49<\/h4>\n<\/p>\n

Lift\u66f2\u7ebf\u662f\u901a\u8fc7\u6bd4\u8f83\u6a21\u578b\u6392\u5e8f\u540e\u7684\u9884\u6d4b\u7ed3\u679c\u4e0e\u968f\u673a\u6392\u5e8f\u7684\u57fa\u51c6\u6765\u8bc4\u4f30\u6a21\u578b\u5728\u4e0d\u540c\u5206\u4f4d\u6570\u4e0a\u7684\u6027\u80fd\u3002\u7eb5\u8f74\u8868\u793a\u63d0\u5347\u5ea6\uff08Lift\uff09\uff0c\u6a2a\u8f74\u8868\u793a\u6837\u672c\u7684\u7d2f\u79ef\u767e\u5206\u6bd4\u3002<\/p>\n<\/p>\n

2\u3001LIFT\u66f2\u7ebf\u7684\u5e94\u7528\u573a\u666f<\/h4>\n<\/p>\n

Lift\u66f2\u7ebf\u901a\u5e38\u7528\u4e8e\u8425\u9500\u3001\u4fe1\u7528\u8bc4\u5206\u548c\u6b3a\u8bc8\u68c0\u6d4b\u7b49\u9886\u57df\u3002\u5b83\u5e2e\u52a9\u51b3\u7b56\u8005\u4e86\u89e3\u6a21\u578b\u5728\u67d0\u4e9b\u7279\u5b9a\u767e\u5206\u6bd4\u7684\u6837\u672c\u4e0b\u7684\u8868\u73b0\uff0c\u4ece\u800c\u505a\u51fa\u66f4\u6709\u9488\u5bf9\u6027\u7684\u51b3\u7b56\u3002<\/p>\n<\/p>\n

\u4e8c\u3001\u4f7f\u7528Python\u5e93\u7ed8\u5236LIFT\u66f2\u7ebf<\/h3>\n<\/p>\n

\u5728Python\u4e2d\uff0c\u6211\u4eec\u53ef\u4ee5\u5229\u7528scikit-learn\u3001matplotlib\u548cpandas\u5e93\u6765\u7ed8\u5236Lift\u66f2\u7ebf\u3002\u4ee5\u4e0b\u662f\u5177\u4f53\u6b65\u9aa4\uff1a<\/p>\n<\/p>\n

1\u3001\u6570\u636e\u51c6\u5907<\/h4>\n<\/p>\n

\u9996\u5148\uff0c\u6211\u4eec\u9700\u8981\u51c6\u5907\u597d\u4e00\u4e2a\u4e8c\u5206\u7c7b\u95ee\u9898\u7684\u6570\u636e\u96c6\u3002\u6570\u636e\u96c6\u5e94\u8be5\u5305\u542b\u771f\u5b9e\u7684\u6807\u7b7e\u548c\u6a21\u578b\u9884\u6d4b\u7684\u6982\u7387\u3002<\/p>\n<\/p>\n

import pandas as pd<\/p>\n
from sklearn.model_selection import trAI<\/a>n_test_split<\/p>\n
from sklearn.ensemble import RandomForestClassifier<\/p>\n
from sklearn.datasets import make_classification<\/p>\n
\u521b\u5efa\u4e00\u4e2a\u793a\u4f8b\u6570\u636e\u96c6<\/strong><\/h2>\n
X, y = make_classification(n_samples=1000, n_features=20, random_state=42)<\/p>\n
\u5206\u5272\u6570\u636e\u96c6<\/strong><\/h2>\n
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=42)<\/p>\n
<\/code><\/pre>\n<\/p>\n
2\u3001\u8bad\u7ec3\u6a21\u578b\u5e76\u83b7\u53d6\u9884\u6d4b\u6982\u7387<\/h4>\n<\/p>\n
\u5728\u8fd9\u4e2a\u6b65\u9aa4\u4e2d\uff0c\u6211\u4eec\u5c06\u4f7f\u7528\u968f\u673a\u68ee\u6797\u5206\u7c7b\u5668\u4f5c\u4e3a\u793a\u4f8b\u6a21\u578b\uff0c\u5e76\u83b7\u53d6\u5176\u9884\u6d4b\u6982\u7387\u3002<\/p>\n<\/p>\n
# \u8bad\u7ec3\u6a21\u578b<\/p>\n
model = RandomForestClassifier(random_state=42)<\/p>\n
model.fit(X_train, y_train)<\/p>\n
\u83b7\u53d6\u9884\u6d4b\u6982\u7387<\/strong><\/h2>\n
y_pred_prob = model.predict_proba(X_test)[:, 1]<\/p>\n
<\/code><\/pre>\n<\/p>\n
3\u3001\u8ba1\u7b97LIFT\u503c<\/h4>\n<\/p>\n
\u901a\u8fc7\u9884\u6d4b\u6982\u7387\uff0c\u6211\u4eec\u53ef\u4ee5\u8ba1\u7b97\u4e0d\u540c\u5206\u4f4d\u6570\u4e0b\u7684Lift\u503c\u3002<\/p>\n<\/p>\n
import numpy as np<\/p>\n
def calculate_lift(y_true, y_pred_prob, num_bins=10):<\/p>\n
    data = pd.DataFrame({'true': y_true, 'pred_prob': y_pred_prob})<\/p>\n
    data.sort_values('pred_prob', ascending=False, inplace=True)<\/p>\n
    data['bin'] = pd.qcut(data['pred_prob'], q=num_bins, duplicates='drop')<\/p>\n
    lift_values = []<\/p>\n
    for bin in data['bin'].unique():<\/p>\n
        bin_data = data[data['bin'] == bin]<\/p>\n
        lift = (bin_data['true'].sum() \/ len(bin_data)) \/ (data['true'].sum() \/ len(data))<\/p>\n
        lift_values.append(lift)<\/p>\n
    return lift_values<\/p>\n
lift_values = calculate_lift(y_test, y_pred_prob)<\/p>\n
<\/code><\/pre>\n<\/p>\n
4\u3001\u7ed8\u5236LIFT\u66f2\u7ebf<\/h4>\n<\/p>\n
\u6700\u540e\uff0c\u6211\u4eec\u4f7f\u7528matplotlib\u5e93\u6765\u7ed8\u5236Lift\u66f2\u7ebf\u3002<\/p>\n<\/p>\n
import matplotlib.pyplot as plt<\/p>\n
def plot_lift_curve(lift_values):<\/p>\n
    plt.figure(figsize=(10, 6))<\/p>\n
    plt.plot(np.arange(1, len(lift_values) + 1), lift_values, marker='o', linestyle='-')<\/p>\n
    plt.title('Lift Curve')<\/p>\n
    plt.xlabel('Quantile')<\/p>\n
    plt.ylabel('Lift')<\/p>\n
    plt.grid(True)<\/p>\n
    plt.show()<\/p>\n
plot_lift_curve(lift_values)<\/p>\n
<\/code><\/pre>\n<\/p>\n
\u4e09\u3001\u4f18\u5316\u548c\u6269\u5c55<\/h3>\n<\/p>\n
\u5728\u7ed8\u5236Lift\u66f2\u7ebf\u7684\u57fa\u7840\u4e0a\uff0c\u6211\u4eec\u53ef\u4ee5\u8fdb\u884c\u4e00\u4e9b\u4f18\u5316\u548c\u6269\u5c55\uff0c\u4ee5\u63d0\u9ad8\u6a21\u578b\u8bc4\u4f30\u7684\u51c6\u786e\u6027\u548c\u53ef\u89c6\u5316\u6548\u679c\u3002<\/p>\n<\/p>\n
1\u3001\u589e\u52a0\u5206\u4f4d\u6570<\/h4>\n<\/p>\n
\u589e\u52a0\u5206\u4f4d\u6570\u53ef\u4ee5\u8ba9Lift\u66f2\u7ebf\u66f4\u52a0\u5e73\u6ed1\uff0c\u5e76\u63d0\u4f9b\u66f4\u7cbe\u7ec6\u7684\u6a21\u578b\u6027\u80fd\u5206\u6790\u3002\u4e0d\u8fc7\uff0c\u8fc7\u591a\u7684\u5206\u4f4d\u6570\u53ef\u80fd\u5bfc\u81f4\u6bcf\u4e2a\u5206\u4f4d\u6570\u5185\u7684\u6570\u636e\u91cf\u4e0d\u8db3\uff0c\u5f71\u54cdLift\u503c\u7684\u7a33\u5b9a\u6027\u3002<\/p>\n<\/p>\n
lift_values_20_bins = calculate_lift(y_test, y_pred_prob, num_bins=20)<\/p>\n
plot_lift_curve(lift_values_20_bins)<\/p>\n
<\/code><\/pre>\n<\/p>\n
2\u3001\u6bd4\u8f83\u591a\u4e2a\u6a21\u578b<\/h4>\n<\/p>\n
\u901a\u8fc7\u7ed8\u5236\u591a\u4e2a\u6a21\u578b\u7684Lift\u66f2\u7ebf\uff0c\u53ef\u4ee5\u76f4\u89c2\u5730\u6bd4\u8f83\u6a21\u578b\u7684\u4f18\u52a3\u3002\u53ea\u9700\u91cd\u590d\u4e0a\u8ff0\u6b65\u9aa4\uff0c\u5c06\u4e0d\u540c\u6a21\u578b\u7684Lift\u66f2\u7ebf\u7ed8\u5236\u5728\u540c\u4e00\u56fe\u4e2d\u3002<\/p>\n<\/p>\n
3\u3001\u4f7f\u7528\u771f\u5b9e\u6570\u636e\u96c6<\/h4>\n<\/p>\n
\u771f\u5b9e\u6570\u636e\u96c6\u901a\u5e38\u66f4\u590d\u6742\uff0c\u56e0\u6b64\u5728\u5b9e\u9645\u5e94\u7528\u4e2d\uff0c\u4f7f\u7528\u771f\u5b9e\u6570\u636e\u96c6\u6765\u7ed8\u5236Lift\u66f2\u7ebf\u80fd\u66f4\u597d\u5730\u53cd\u6620\u6a21\u578b\u7684\u6027\u80fd\u3002<\/p>\n<\/p>\n
\u56db\u3001\u5728\u5b9e\u8df5\u4e2d\u4f7f\u7528LIFT\u66f2\u7ebf<\/h3>\n<\/p>\n
\u7406\u89e3\u5e76\u6b63\u786e\u4f7f\u7528Lift\u66f2\u7ebf\u80fd\u5e2e\u52a9\u6211\u4eec\u66f4\u597d\u5730\u8fdb\u884c\u6a21\u578b\u8bc4\u4f30\u4e0e\u9009\u62e9\u3002\u5728\u5b9e\u8df5\u4e2d\uff0c\u6211\u4eec\u9700\u8981\u6839\u636e\u5177\u4f53\u95ee\u9898\u9009\u62e9\u5408\u9002\u7684\u5206\u4f4d\u6570\u3001\u6a21\u578b\u548c\u6570\u636e\u96c6\u3002<\/p>\n<\/p>\n
1\u3001\u7ed3\u5408\u4e1a\u52a1\u9700\u6c42<\/h4>\n<\/p>\n
\u5728\u8425\u9500\u9886\u57df\uff0cLift\u66f2\u7ebf\u53ef\u4ee5\u5e2e\u52a9\u6211\u4eec\u8bc6\u522b\u9ad8\u6f5c\u529b\u5ba2\u6237\u7fa4\u4f53\u3002\u5728\u6b3a\u8bc8\u68c0\u6d4b\u4e2d\uff0c\u5b83\u80fd\u5e2e\u52a9\u6211\u4eec\u627e\u51fa\u6700\u53ef\u80fd\u7684\u6b3a\u8bc8\u884c\u4e3a\u3002\u7ed3\u5408\u4e1a\u52a1\u9700\u6c42\uff0c\u53ef\u4ee5\u66f4\u6709\u9488\u5bf9\u6027\u5730\u5229\u7528Lift\u66f2\u7ebf\u3002<\/p>\n<\/p>\n
2\u3001\u4e0e\u5176\u4ed6\u8bc4\u4f30\u6307\u6807\u7ed3\u5408<\/h4>\n<\/p>\n
\u5c3d\u7ba1Lift\u66f2\u7ebf\u662f\u4e00\u4e2a\u5f3a\u5927\u7684\u5de5\u5177\uff0c\u4f46\u5b83\u4e0d\u5e94\u5355\u72ec\u7528\u4e8e\u6a21\u578b\u8bc4\u4f30\u3002\u7ed3\u5408\u5176\u4ed6\u6307\u6807\uff0c\u5982AUC-ROC\u3001F1-score\u7b49\uff0c\u53ef\u4ee5\u5168\u9762\u5206\u6790\u6a21\u578b\u7684\u6027\u80fd\u3002<\/p>\n<\/p>\n
\u901a\u8fc7\u4ee5\u4e0a\u6b65\u9aa4\uff0c\u6211\u4eec\u80fd\u591f\u5229\u7528Python\u7ed8\u5236Lift\u66f2\u7ebf\uff0c\u5e76\u5728\u5b9e\u8df5\u4e2d\u5e94\u7528\u5176\u8bc4\u4f30\u6a21\u578b\u6027\u80fd\u7684\u80fd\u529b\u3002\u5e0c\u671b\u8fd9\u7bc7\u6587\u7ae0\u5bf9\u60a8\u7406\u89e3Lift\u66f2\u7ebf\u53ca\u5176\u5728Python\u4e2d\u7684\u5b9e\u73b0\u6709\u6240\u5e2e\u52a9\u3002<\/p>\n<\/p>\n
\u76f8\u5173\u95ee\u7b54FAQs\uff1a<\/strong><\/h2>\n
 \u5982\u4f55\u7528Python\u7ed8\u5236Lift\u66f2\u7ebf\uff1f<\/strong>
\u5728Python\u4e2d\uff0c\u7ed8\u5236Lift\u66f2\u7ebf\u901a\u5e38\u9700\u8981\u4f7f\u7528\u6570\u636e\u5206\u6790\u5e93\u548c\u53ef\u89c6\u5316\u5e93\u3002\u53ef\u4ee5\u5229\u7528pandas<\/code>\u6765\u5904\u7406\u6570\u636e\uff0cscikit-learn<\/code>\u6765\u8fdb\u884c\u6a21\u578b\u8bc4\u4f30\uff0c\u6700\u540e\u7528matplotlib<\/code>\u6216seaborn<\/code>\u6765\u7ed8\u5236Lift\u66f2\u7ebf\u3002\u9996\u5148\uff0c\u786e\u4fdd\u4f60\u6709\u771f\u5b9e\u6807\u7b7e\u548c\u9884\u6d4b\u5f97\u5206\uff0c\u63a5\u7740\u53ef\u4ee5\u6309\u7167\u4e00\u5b9a\u7684\u6bd4\u4f8b\u8ba1\u7b97Lift\u503c\uff0c\u7136\u540e\u7ed8\u5236\u56fe\u5f62\u3002<\/p>\n
Lift\u66f2\u7ebf\u7684\u610f\u4e49\u662f\u4ec0\u4e48\uff1f<\/strong>
Lift\u66f2\u7ebf\u7528\u4e8e\u8bc4\u4f30\u5206\u7c7b\u6a21\u578b\u7684\u6548\u679c\uff0c\u7279\u522b\u662f\u5728\u4e0d\u5e73\u8861\u6570\u636e\u96c6\u4e2d\u3002\u901a\u8fc7\u8ba1\u7b97\u4e0d\u540c\u9608\u503c\u4e0b\u7684\u63d0\u5347\u7387\uff0cLift\u66f2\u7ebf\u53ef\u4ee5\u5e2e\u52a9\u4f60\u4e86\u89e3\u6a21\u578b\u5728\u6b63\u6837\u672c\u9884\u6d4b\u4e2d\u7684\u4f18\u8d8a\u6027\u3002Lift\u503c\u8d8a\u9ad8\uff0c\u8bf4\u660e\u6a21\u578b\u7684\u9884\u6d4b\u80fd\u529b\u8d8a\u5f3a\u3002<\/p>\n
\u5728\u7ed8\u5236Lift\u66f2\u7ebf\u65f6\uff0c\u6709\u54ea\u4e9b\u5e38\u89c1\u7684\u9519\u8bef\u9700\u8981\u907f\u514d\uff1f<\/strong>
\u5728\u7ed8\u5236Lift\u66f2\u7ebf\u65f6\uff0c\u5e38\u89c1\u7684\u9519\u8bef\u5305\u62ec\u672a\u6b63\u786e\u6392\u5e8f\u9884\u6d4b\u6982\u7387\u3001\u5ffd\u89c6\u7c7b\u522b\u4e0d\u5e73\u8861\u5bf9\u7ed3\u679c\u7684\u5f71\u54cd\u3001\u4ee5\u53ca\u5728\u7ed8\u56fe\u65f6\u672a\u6807\u793a\u51fa\u57fa\u7ebf\u6216\u968f\u673a\u6a21\u578b\u7684\u8868\u73b0\u3002\u786e\u4fdd\u6570\u636e\u6e05\u6d17\u548c\u9884\u5904\u7406\u6b63\u786e\uff0c\u4ee5\u53ca\u5728\u56fe\u4e2d\u6e05\u6670\u6807\u793a\u4e0d\u540c\u7684\u66f2\u7ebf\uff0c\u6709\u52a9\u4e8e\u66f4\u51c6\u786e\u5730\u89e3\u8bfb\u7ed3\u679c\u3002<\/p>\n","protected":false},"excerpt":{"rendered":"Python\u4e2d\u7ed8\u5236Lift\u66f2\u7ebf\u53ef\u4ee5\u901a\u8fc7\u4f7f\u7528scikit-learn\u5e93\u3001matplotlib\u5e93\u548cpandas\u5e93\u5b9e […]","protected":false},"author":3,"featured_media":979805,"comment_status":"closed","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":""},"categories":[37],"tags":[],"acf":[],"_links":{"self":[{"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/posts\/979801"}],"collection":[{"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/users\/3"}],"replies":[{"embeddable":true,"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/comments?post=979801"}],"version-history":[{"count":"1","href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/posts\/979801\/revisions"}],"predecessor-version":[{"id":979807,"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/posts\/979801\/revisions\/979807"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/media\/979805"}],"wp:attachment":[{"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/media?parent=979801"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/categories?post=979801"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/tags?post=979801"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}