Computer scienceData scienceMachine learningEnsemble learning

Adaboost

Theory

Gradboost vs Adaboost

Report a typo

Match the statements that hold true for Adaboost or Gradboost.

Hint

Carefully review A comparison between Gradboost and Adaboost section of the theory.

Choose one or more options for each row
GradboostAdaboost
The weak learners are fit sequentially
More sensitive to the outliers
Can use any diffirentiable loss function
Mainly uses the exponential loss for classification, but could be formulated outside the context of loss minimization
Assigns weights to each training sample, and each iteration places extra emphasis on the misclassified samples
The weighted majority vote is used to produce the final prediction
The output is an additive models of multiple weak learners

Create a free account to access the full topic