好东西

In R, there is a package called "RandomForest", you can use it to realize RandomForest.  By the way, all most prediction algorithm such as Logistic , NN, Decision Tree(Boosting, Bagging,RamdomForest) can handle overfitting. You need to split your data before you build your model. Training data set for model building, validation dataset for model tuning. Then you can calculate fit statistics( misclassification rate, Average Square Error, ROC Index etc) on validation dataset to optimize the model complexity. Therefore, data splitting is necessary no matter which method you use.

楼上的，那你知不知道怎么实现。

你知道怎么写DECISION TREE的程序吗

如果知道,一个循环就是RANDOM FOREST的程序

我想问你要用FORTRAN写的RANDOM FOREST的程序,不知道是否愿意和我分享一下

QQ:510521607

路過 學習一下