看到文献中关于两者的描述“As CART's principal weakness is its instability with regard to changes in the learning sample (Breiman et al., 1984), we also considered Random Forests models (Breiman 2001), an ensemble learning technique that avoids this problem and has shown to out-perform CART with respect to prediction error (Breiman, 2001). Furthermore, this technique presents lower error rates compared to other statistical tools in ecological studies (e.g.: Cutler et al., 2007; Cirsci et al., 2012).”
试了rpart包,能画出节点图。接着试了randomForest包,却找不到画节点图的地方。
看到上述文献中描述到,rpart包结果更容易理解;也看了randomForest包的介绍,感觉确实randomForest包不能画节点图。
但是我还是不确定,既然节点图更好理好,为什么randomForest不提供这个功能呢?
类似于下面的节点图: