# 13-3 bagging和pasting

虽然有很多机器学习的算法，但从投票的角度看，仍然不够多\
创建更多的子模型，集成更多子模型的意见\
子模型之间不能一致，子模型之间要有差异性\
如果创建差异性？

解决方法：\
每个子模型只看样本数据的一部分。\
每个子模型不太需要太高的准确率。只要子模型足够多，准确率就会提高。\
例如500个子模型，每个子模型的准确率是60%，最终准确率能达到99.9%

取样方法：

* 放回取样 bagging（bootstrap）   
* 不放回取样 pasting   

  bagging更常用，优点：  
* 没有那么依赖随机  
* 数据量要求没那么高  

使用13-1中的数据

```python
from sklearn.tree import DecisionTreeClassifier
from sklearn.ensemble import BaggingClassifier

bagging_clf = BaggingClassifier(DecisionTreeClassifier(),n_estimators=500, max_samples=100, bootstrap=True)
# 决策树这种非参数的算法更容易产生差异较大的子模型
# 所有集成学习如果要集成成百上千个子模型，通常首先决策树
# n_estimators：子模型数
# max_samples：每个子模型看的样本树
# bootstrap：放回取样

bagging_clf.fit(X_train, y_train)
bagging_clf.score(X_test, y_test)
```

输出：0.912

```
bagging_clf2 = BaggingClassifier(DecisionTreeClassifier(),n_estimators=1000, max_samples=100, bootstrap=True)
bagging_clf2.fit(X_train, y_train)
bagging_clf2.score(X_test, y_test)
```

输出：0.92


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://windmising.gitbook.io/liu-yu-bo-play-with-machine-learning/src/chapter13-2/13-3.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.