Preprocessing the TITANIC dataset and using predictive models