Bagging (Bootstrap Aggregating) is an ensemble method that trains multiple models on different bootstrap samples of the dataset and averages their predictions. This approach reduces variance and often improves stability and accuracy compared to a single model. Random Forests are a well-known example of a bagging technique.