AI 용어집
인공지능 완전 사전
Weighting Scheme
Mechanism for assigning weights to training instances and models in boosting, where weights increase for misclassified instances at each iteration.
Sequential Learning
Learning paradigm where each model is trained sequentially considering the performance of previous models to correct their systematic errors.
Residual Fitting
Gradient boosting technique where each new model is trained to predict the residuals (errors) of the previous ensemble model rather than the original target.
Tree Boosting
Specific application of boosting using decision trees as weak learners, particularly effective for non-linear regression and classification problems.
Regularization in Boosting
Set of techniques (shrinkage, subsampling, L1/L2 penalties) controlling the complexity of the boosting model to prevent overfitting and improve generalization.
Loss Function Optimization
Mathematical process iteratively minimizing a specific loss function (exponential, logistic, huber) to guide the sequential learning of boosting models.
Out-of-Bag Estimation
Evaluation method for boosting models using instances not selected in stochastic subsampling as an internal validation set.