Information Gain and Mutual Information

Information gain is routinely used to determine how a decision tree splits but it can also be used to determine which features best predict the outcome for each algorithm.