QUESTION 9

1. When we have a small amount of training data, which method shall we use to validate the model?

   		Training and validating on the full set of training data
   		Holding back 20% of the data from the training of the model, and use this holdout set to check the model performance
   		Use cross-validation by training several machine learning models on subsets of the available training data and evaluating them on the complementary subset of the data
   		Use the testing dataset to validate the model