增强回归树模型

时间:2019-07-07 05:07:06

标签: tree regression gbm

如何在代码中放置复制?

如何根据结果确定R平方?

如何获得超过1000棵树,但我的lr已经为0.001,这是一个问题吗?

例如:我尝试过tr 1,2,3,4,结果仍然小于1000。 以下是我使用的代码:

`tesbrt <- gbm.step(data=shrub, gbm.x = 7:11, gbm.y = 13,
                        family = "gaussian", tree.complexity = 1,
                        learning.rate = 0.001, bag.fraction = 0.6)
angaus.simp <- gbm.simplify(tesbrt, n.drops = 5)
gbm.plot(tesbrt, n.plots=5, plot.layout=c(2, 3), write.title = FALSE)`

我得到的结果:

mean total deviance = 2.198
mean residual deviance = 1.967

estimated cv deviance = 2.198 ; se = 0.287

training data correlation = 0.525
cv correlation = 0.103 ; se = 0.076

elapsed time - 0.25 minutes

0 个答案:

没有答案