The cross_validation function was not working on learning.py. Running the compare function, which is responsible for comparing multiple algorithms against their cross validation score would generate an error. With that said, this MR does the following:

• Update cross_validation function to resemble more the book pseudocode
• Update cross_validation_wrapper to resemble more the book pseudocode
• Add tests to function train_and_test
• Update print_table function on utils.py to better handle the cross validation output.
• Add dummy size variable to some learner implementations, in order to maintain cross_validation similiarity with the book algorithm. This should be updated in the future, in order for the size parameter to be a way to better configure a learner.