Chapter VI
Machine Learning
73
VI.3.4. Random Forest
Random Forest is a powerful supervised learning algorithm utilized for various tasks such as
classification, regression, and more. It functions by constructing multiple decision trees during
the training phase. The algorithm's essence lies in creating an ensemble of decision trees Figure
VI-2, forming a forest. This implies that numerous trees contribute to the forest's creation,
serving as the fundamental principle behind the random forest algorithm. The prediction is
determined by aggregating the predictions of all the trees.
Figure VI-2 The random forest algorithm relies on multiple decision trees that are all trained slightly
differently.
VI.3.5. Gradient Boosting
Gradient Boosting Machines, or GBMs for short, are classified as sequential ensemble models,
meaning they build predictive models by combining multiple weak learners in a sequential
manner. GBMs excel in various tasks such as regression, classification, and ranking, and are
highly favored by data scientists and practitioners due to their ability to boost model accuracy
and performance.
Figure VI-3 Flow diagram of gradient boosting machine learning method.