 What are the different ways to perform data validation? Validation is required in order to truly understand how your model will perform in the real world and two primary ways of doing this is the first is hold out validation where we'll take our data set and we'll use a chunk of it for training and the remaining chunks specifically for Testing and let's say that we get a mean squared error, which we compute only from the test chunk Another version is cross validation where we divide the data into multiple chunks We train with the red chunks and test with the green chunks to generate one error and we do this selecting different chunks for testing in order to generate different errors and Once we have all the errors we simply take the average to get the overall cross validation error