可变重要性图

发布于 2025-01-21 23:23:46 字数 974 浏览 0 评论 0原文

我正在从随机森林回归中研究可变的重要性图，并希望使用VIP软件包将可变标签应用于Y轴而不是隐秘的变量名称，以易于解释。

如何修改代码以在Y轴上添加标签？

我当前的代码：

library(randomForest)
library(vip)
model1 <- randomForest(y ~ .,
                      data=train,
                      mtry = 8,
                      ntree=500,
                      import[enter image description here][1]ance=TRUE,
                      proximity=TRUE)

plot_rf_1 <- vip(model1, 
                 num_features = 6,
                 include_type = TRUE,
                 geom = "point",
                 horizontal = TRUE,
                 mapping = aes_string(fill="var.labels"),
                 aesthetics = list(color = "blue", shape = 16, size = 3)) +
                theme_light()
grid.arrange(plot_rf_1, nrow = 1)```

可变重要性图：

原文

I'm working on variable importance plot from random forest regression and want to apply variable labels to y-axis instead of cryptic variable names using the VIP package for ease of interpretation.

How to modify the code to add labels on y axis?

My current code:

library(randomForest)
library(vip)
model1 <- randomForest(y ~ .,
                      data=train,
                      mtry = 8,
                      ntree=500,
                      import[enter image description here][1]ance=TRUE,
                      proximity=TRUE)

plot_rf_1 <- vip(model1, 
                 num_features = 6,
                 include_type = TRUE,
                 geom = "point",
                 horizontal = TRUE,
                 mapping = aes_string(fill="var.labels"),
                 aesthetics = list(color = "blue", shape = 16, size = 3)) +
                theme_light()
grid.arrange(plot_rf_1, nrow = 1)```

Variable importance plot: