How it works...

The random forest was developed by Leo Breiman (University of California, Berkeley, USA) based on the use of classification trees. He has extended the classification tree technique by integrating it into a Monte Carlo simulation procedure and named it random forest. It is based on the creation of a large set of tree classifiers, each of which is proposed to classify a single instance, wherein some features have been evaluated. Comparing the classification proposals provided by each tree in the forest shows the class to which to attribute the request: it is the one that received the most votes.