Given a dataset with multiple variables, how would you determine which are most influential in predicting an outcome?