Iris inverse formula in rxNeuralNet

I'm trying to understand a little of the mechanics behind how "R" treats factors as predictor variables. None of what I write below is probably good practice, but out of sheer curiosity, any thoughts would be appreciated. The standard Iris dataset in R has columns: Sepal.Length, Sepal.Width, Petal.Length, Petal.Width, Species, the last of which is a factor. The standard relation to this dataset is to demonstrate classification algorithms, for example using a neural network or tree, where

rnn1 <- rxNeuralNet(Species~Sepal.Length+Sepal.Width+...,data = iris, numHiddenNodes = 100, numIterations = 1000,type = "classification")

      

        
        
        
      

    

I decided to see what happens if you change this to:

rnn2 <- rnn1 <- rxNeuralNet(Petal.Width~Sepal.Length+Sepal.Width+Species,data = iris, numHiddenNodes = 100, numIterations = 1000,type = "regression")

Then I created my test dataframe:

df1 <- data.frame(Petal.Width=5,Sepal.Length=12,Sepal.Width=3,Species="setosa",Petal.Length=3)

      

        
        
        
      

    

rxPredict () then gives me a score of 0.6058862 for the "setosa" species. But oddly enough, and this is my question, I can put any "string" I want for the species, and I still get a prediction. I put Species="Jack"

, and rxPredict now gives an estimate of 1.545223. This is weird because in standard-R it will generate a factor error if you try to predict any factors that were not in the original dataset.

Any ideas?

Thank.