For the Seeds data example in lab notes, how many clusters will you choose as the final model? Why? Please show your work.