Importance of fields in the dataset in randomforest

Question

Hi. I'm using randomforest method to make a prediction of testdata. Is there a way to display the value or order of fields?

Answer 1

Use Function importance() which defines the importance or order of fields in a decision tree.

The functions return a data frame with field names, accuracy for each prediction and accuracy is displayed.

Ex: Importance(randomforest_model).

mtcars.rf <- randomForest(mpg ~ ., data=mtcars)
> importance(mtcars.rf)
       %IncMSE IncNodePurity
cyl  17.311478     180.55303
disp 20.064044     250.82581
hp   17.506333     187.55177
drat  7.041252      61.64077
wt   18.566081     248.37960
qsec  3.518307      32.91833
vs    5.418713      26.88210
am    5.296174      17.83260
gear  3.807433      19.50622
carb  7.725077      28.68420