简体   繁体   English

如何为随机森林回归同时调整 mtry 和树数?

[英]How to tune mtry and number of trees simultaneously for a Random Forest Regression?

I am trying to tune parameters for a Random Forest using caret and method ranger .我正在尝试使用caret和方法ranger调整随机森林的参数。 I have seen codes for tuning mtry using tuneGrid .我已经看到了使用tuneGrid调整mtry的代码。 And then using the resulted mtry to run loops and tune the number of trees ( num.tree ).然后使用生成的mtry运行循环并调整树的数量( num.tree )。 However, I would like to know if it is possible to tune them both at the same time, to find out the best model between all possible combinations.但是,我想知道是否可以同时调整它们,以找出所有可能组合之间的最佳模型。 I do not want to keep one argument constant and tune the other one, but both at the same time.我不想让一个论点保持不变并调整另一个论点,但要同时调整两者。 Is there any way?有什么办法吗?

You cannot tune ntree as part of a tuneGrid for Random Forest in caret ;您不能在caret中将ntree作为随机森林的tuneGrid的一部分进行调整; only mtry , splitrule and min.node.size - see the tuning parameters for each model type here: https://topepo.github.io/caret/available-models.html只有mtrysplitrulemin.node.size - 在此处查看每种模型类型的调整参数: https ://topepo.github.io/caret/available-models.html

ntree can only be specified in train . ntree只能在train中指定。

声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.

 
粤ICP备18138465号  © 2020-2024 STACKOOM.COM