Short for Adaptive Boosting, this ensemble classifier can improve the performance of an otherwise weak classifier by focusing more attention on samples that are harder to classify. It builds an additive model where, at each stage, a new learner is trained and given an influence score inversely proportional to the loss it incurs at that epoch.
Note: The default base learner is a Classification Tree with a max depth of 1 i.e a Decision Stump.
Data Type Compatibility: Depends on base learner
|1||base||ClassificationTree||Learner||The base weak classifier to be boosted.|
|2||rate||1.0||float||The learning rate of the ensemble i.e. the shrinkage applied to each step.|
|3||ratio||0.8||float||The ratio of samples to subsample from the training set to train each weak learner.|
|4||estimators||100||int||The maximum number of weak learners to train in the ensemble.|
|5||min change||1e-4||float||The minimum change in the training loss necessary to continue training.|
|6||window||5||int||The number of epochs without improvement in the training loss to wait before considering an early stop.|
use Rubix\ML\Classifiers\AdaBoost; use Rubix\ML\Classifiers\ExtraTreeClassifier; $estimator = new AdaBoost(new ExtraTreeClassifier(3), 0.1, 0.5, 200, 1e-3, 10);
Return the loss at each epoch from the last training session:
public steps() : float|null
Return the influence scores for each classifier in the ensemble:
public influences() : float
- Y. Freund et al. (1996). A Decision-theoretic Generalization of On-line Learning and an Application to Boosting.
- J. Zhu et al. (2006). Multi-class AdaBoost.