0

I am using this code:

library(datasets)
library(quantregForest)
library(dplyr)
library(ggplot2)
library(randomForest)

Xtrain <- iris %>%
    select(
        Sepal.Length
        , Petal.Length
    )

Ytrain <- iris$Petal.Width

model <- quantregForest(
    x = Xtrain
    , y = Ytrain
)

plot.partial = function() {
    partialPlot(model, Xtrain, x.var = "Sepal.Length")
    partialPlot(model, Xtrain, x.var = "Petal.Length")
}

class(model) <- "randomForest"
plot.partial()

What does the partial dependence plot actually show? Is this the mean prediction (rather than for example the median)?

cs0815
  • 1,294
  • 18
  • 30
  • 2
    Possible duplicate of [R: What do I see in partial dependence plots of gbm and RandomForest?](https://stats.stackexchange.com/questions/92150/r-what-do-i-see-in-partial-dependence-plots-of-gbm-and-randomforest) – paoloeusebi Nov 29 '18 at 17:15
  • 1
    I was aware of this but I would not agree with your statement. My question was more related to the fact the model returns quantiles . . – cs0815 Nov 29 '18 at 18:02

0 Answers0