Step 5 – Splitting the dataset into test and train datasets