查询随机森林中的变量选择

Question

我对随机森林中的变量选择有一点疑问。我知道它选择＆＃34; m＆＃34;随机变量出现在＆＃34; M＆＃34;用于拆分的变量并保持值（m）始终保持不变。

我的问题是为什么这些m变量在每个节点都不相同。它背后的原因是什么？有人可以帮忙吗。

谢谢，

Answer 1

Fact that it is using different set (randomly chosen) of m features for each tree is actually advantage for RF. That way final model is more robust and accurate. It also helps in identifying which features are contributing most and have best predictive power.
btw that's why it is called Random Forest after all...