Labelling Variable Importance Plot

Hi All,

I've been following a published script on conditional variable importance, here:Be Aware of Bias in RF Variable Importance Metrics | R-bloggers

Yet, I've noticed when I follow this particular script that I am unable to obtain variable names on the vertical axis. Instead, they are identified by their column number.

I wondered if someone knew of a way around this? Would be appreciated

library(tidyverse)
library(caret)
library(ranger)
library(ggplot2)
library(party)

rf3 <- cforest( score ~ ., data = readingSkills, control = cforest_unbiased(mtry = 2, ntree = 500)

imp <- rf %>%
    varimp(conditional = TRUE) %>% 
    as_tibble() %>% 
    rownames_to_column("Feature") %>% 
    rename(Importance = value)
  
  ggplot(imp, aes(x = reorder(Feature, Importance), y = Importance)) +
       geom_bar(stat = "identity", fill = "#53cfff", width = 0.65) +
       coord_flip() + 
       theme_light(base_size = 20) +
       theme(axis.title.x = element_text(size = 15, color = "black"),
             axis.title.y = element_blank(),
             axis.text.x  = element_text(size = 15, color = "black"),
             axis.text.y  = element_text(size = 15, color = "black"))

Screenshot 2021-11-18 at 17.55.16

WARNING. THE SCRIPT TAKES AROUND 15-20 MINS

I reduced the ntree down from 500 to 100 to reduce waiting time... I found I had to add a missing bracket from the end there.
also when you make imp , there is no rf to make it from, rather rf3.

The problem you had was the varimp result gives a named numeric vector, as_tibble default behaviour wont capture that as nicely as other more appropriate tooling, such as dplyr::enframe.

library(tidyverse)
library(caret)
library(ranger)
library(party)

rf3 <- cforest( score ~ ., data = readingSkills, control = cforest_unbiased(mtry = 2, ntree = 100))
                
                imp <- rf3 %>%
                  varimp(conditional = TRUE) %>% 
                  enframe() %>% 
                  rename(Feature = name,
                         Importance = value)
                
                ggplot(imp, aes(x = reorder(Feature, Importance), y = Importance)) +
                  geom_bar(stat = "identity", fill = "#53cfff", width = 0.65) +
                  coord_flip() + 
                  theme_light(base_size = 20) +
                  theme(axis.title.x = element_text(size = 15, color = "black"),
                        axis.title.y = element_blank(),
                        axis.text.x  = element_text(size = 15, color = "black"),
                        axis.text.y  = element_text(size = 15, color = "black"))

Thank you @nirgrahamuk . That worked perfectly for me. Thanks for your help

This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.

If you have a query related to it or one of the replies, start a new topic and refer back with a link.