重采样：SMOTE 和欠采样
重新平衡训练集
重采样：SMOTE 和欠采样 是 CoddyKit 上的免费 Data Science Academy 课时。 这是第 2 节课，共 4 节。 你可以在下方免费阅读本课时的完整内容 — 然后在浏览器中使用内置代码编辑器和全天候 AI 导师进行实践。 这是 Data Science Academy 学习路径的一部分，你的进度在网页和 CoddyKit 应用中同步。 Data Science Academy 课程共包含 4 节课。
本课时的部分内容尚未翻译，以英文显示。
Rebalance the Training Set
One way to help a model see the rare class is to resample: change how many examples of each class it trains on.
Two Directions to Balance
You can add more of the minority (oversample) or trim the majority (undersample). Both aim to even out the class counts.
Naive Oversampling
The simplest oversample just copies minority rows until counts match. It works, but those exact duplicates can make a model memorize them.
Enter SMOTE
SMOTE creates synthetic minority examples instead of copies. It invents new, plausible points between real minority neighbors.
How SMOTE Builds Points
SMOTE picks a minority row, finds a nearby minority neighbor, and places a fresh point somewhere on the line between them.
SMOTE in Code
The imbalanced-learn library makes SMOTE a two-line affair on your training features and labels.
from imblearn.over_sampling import SMOTE
X_res, y_res = SMOTE().fit_resample(X_train, y_train)Undersampling the Majority
Undersampling drops majority rows until balance returns. It trains faster but throws away data, which can cost real signal.
When to Pick Which
Oversample when data is scarce and every row matters. Undersample when the majority is huge and you can spare some rows for speed.
The Cardinal Rule
Resample only the training set. Touch the test set and your scores become fantasy, since real data is never rebalanced for you.
Fit on Train Only
SMOTE must learn from training rows alone. Leaking test rows into resampling inflates results and quietly causes data leakage. ⚠️
Resampling Is Not a Cure-All
Balanced counts help, but they are not magic. Pair resampling with the right metrics and sometimes class weights for the best results.
Quick Check
What makes SMOTE different from plain oversampling?
Recap
Resampling balances the training set by oversampling (SMOTE makes synthetic rows) or undersampling. Never resample the test set. 🎯
常见问题解答
「重采样：SMOTE 和欠采样」课时是免费的吗？
是的 — 「重采样：SMOTE 和欠采样」的完整文本可在网页上免费阅读。要进行交互式练习（内置代码编辑器和全天候 AI 导师）并解锁 Data Science Academy 课程的其余内容，请升级到 CoddyKit PRO。 Data Science Academy 课程共包含 4 节课。
「重采样：SMOTE 和欠采样」这节课中我会学到什么？
重新平衡训练集 你通过在浏览器中直接运行的动手代码来练习 Data Science Academy，全天候 AI 导师会在你学习这节课的过程中回答你的问题。
学习 Data Science Academy 需要有经验吗？
无需任何先前经验。CoddyKit 上的 Data Science Academy 课程适合初学者到高级学习者，你可以从这里开始或从头开始，按照自己的节奏学习。 这是第 2 节课，共 4 节。
「重采样：SMOTE 和欠采样」课时需要多长时间？
大多数 CoddyKit 课程大约需要 5–10 分钟。每节课都很精短且互动，所以你能稳步进步，并在网页和应用中从离开的地方继续。
我能在这节 Data Science Academy 课中编写并运行代码吗？
能。每节 Data Science Academy 课都包含内置代码编辑器，你可以在浏览器中直接编写并运行真实代码，并获得即时 AI 反馈 — 无需本地设置。
此课程中的所有课时
- 类别不平衡时准确率为何会误导
- 重采样：SMOTE 和欠采样
- 类别权重和阈值
- 为稀有事件选择指标