What ensemble learning method builds multiple decision trees and combines their predictions to enhance accuracy and robustness?

The method that builds multiple decision trees and combines their predictions to enhance accuracy and robustness is known as Random Forest. This ensemble learning technique operates by creating a collection of decision trees during training time and aggregating their results to improve the overall predictive performance.

Random Forest leverages the concept of "bagging," where subsets of the training data are used to build each individual tree. By averaging the predictions of these multiple trees — or through majority voting for classification tasks — Random Forest mitigates overfitting and increases generalization to unseen data. This robustness is achieved because different trees will capture different patterns and relationships within the data, and when combined, they yield a more stable and reliable prediction.

Other methods listed, such as Gradient Boosting, also involve multiple trees, but they build them sequentially and focus on correcting the errors of previous trees rather than combining them simultaneously as Random Forest does. Support Vector Machine operates on a different principle, utilizing hyperplanes for classification, while a single Decision Tree Classifier does not capitalize on the benefits gained from ensemble methods.