How to obtain class labels of xgboost predictions? - r

Question

I am using xgboost similar to the following example, where I "recode" some numeric value to a (numeric) value in 0,1,2 indicating class labels. Note that I did not convert this to a factor variable.
Then I use xgboost to fit a model and produce predictions.
library(xgboost)
iris$Species <- runif(nrow(iris))
recode <- function(x){
if(x >= 0 & x <= 0.33){
x <- 0
} else if(x > 0.33 & x <= 0.66){
x <- 1
} else if(x > 0.66){
x <- 2
}
}
train <- xgb.DMatrix(data = as.matrix(iris[,-5]),
label = sapply(iris$Species, FUN = recode))
bst <- xgboost(data = train,
max_depth = 4, eta = 0.5, nrounds = 10,
objective = "multi:softprob",
num_class = 3)
pred <- predict(bst, as.matrix(iris[, -5]), reshape = TRUE)
str(pred)
Is there away to obtain the column labels of the matrix of predictions? Or can I be sure that they are ordered according to the numeric values to which I recoded the input?

Answer 1

The columns follow the same order as your label, so it's 0,1 and 2. To be sure, you can do a confusion matrix to check whether you are predicting it correctly:
library(xgboost)
set.seed(100)
iris$Species <- runif(nrow(iris))
train <- xgb.DMatrix(data = as.matrix(iris[,-5]),
label = sapply(iris$Species, FUN = recode))
bst <- xgboost(data = train,
max_depth = 4, eta = 0.5, nrounds = 10,
objective = "multi:softprob",
num_class = 3)
pred <- predict(bst, as.matrix(iris[, -5]), reshape = TRUE)
# which.max tells you which column is most probable
# we convert them back to 0-2, assuming column 1 corresponds to 0
predicted = apply(pred,1,which.max)-1
actual = sapply(iris$Species,recode)
table(predicted,actual)
The results are:
actual
predicted 0 1 2
0 36 2 2
1 4 48 4
2 6 3 45
So most of those predicted to be 0,1 or 2 follows the highest probable class predicted.
Or if you use caret:
caret::confusionMatrix(factor(predicted,levels=1:3),factor(actual,levels=1:3))