Adversarial Learning in Statistical Classification: A Comprehensive Review of Defenses Against Attacks