Before it makes a decision, the method combines K nearest neighbors, support vector machine, and decision tree learning. Accuracy is reportedly up to 89%.