Building Random Forest on a data-set comprising of missing NA values

Question

I have a modified "iris" dataset comprising of missing values:

iris1$Sepal.Length[c(1,3,57,103)]<-NA

 and i want to build the "Random Forest" algorithm on top of it:

randomForest(Species~Sepal.Length,data=iris1)

But i get this error:

Error in na.fail.default(list(Species = c(1L, 1L, 1L, 1L, 1L, 1L, 1L,  : missing values in object

Is there a way i can build the "random forest" algorithm on top of it?

Answer 1

You have two options, either impute the missing values or omit the missing values.

If you want to impute the missing values in the predictor data, you can use rfImpute() function from randomForest package.

You can run the below command which will impute the missing values in the predictor data:

rfImpute(Species~.,data=iris1)->iris1

Now you can go ahead and use the randomForest function to build the "random Forest" algorithm on top of the iris1 dataset:

randomForest(Species~Sepal.Length,data=iris1)

If there are only few missing values in your data-set you can go ahead and remove them using na.omit() function:

na.omit(iris1)->iris1

After removing the missing values, you can go ahead and build the randomForest function on top of the "iris1" dataset:

randomForest(Species~Sepal.Length,data=iris1)