Question

True/ False: Boosting can combine the strength of several classification procedures (e.g.,

logistic, LDA, SVM) to create a linearly weighted sum classifier.

Answer 1

FALSE Boosting is a technique in which a strong classifier is built from a number of weak classifiers in a linear fashion. These weak classifiers are not formed from several classification procedures like logistics, LDA, SVM but formed on the basis of the changes in sampling distribution of the training data set at each iteration which is used to correct the errors in previous iteration of models. Hence, there are many iterations on the same training dataset by adjusting the weights.